Medijska pismenost građana Srbije, Izveštaj anketnog istraživanja

Project “Strengthening Freedom of Expression and Freedom of Media in Serbia”, OSCE Mission to Serbia

MEDIA LITERACY OF SERBIA’S CITIZENS Survey Research Report Belgrade, 2025

Research Team: Marijana Matović, Faculty of Political Science – University of Belgrade; Chief Researchers Ana Milojević Faculty of Political Science – University of Belgrade and Snežana Milin Perković; Aleksandar Tomašević, Institute of Physics, University of Belgrade; Slavica Josifović and Milena Jović.

This project is funded by the European Union

Disclaimer: This research was published thanks to the financial support of the European Union. The views herein expressed are solely those of the authors and contributors and do not necessarily reflect the official position of the OSCE Mission to Serbia and can in no way be taken to reflect the views of European Union.

1 Contents

Research highlights ................................................................................................................................................4 About the Research...................................................................................................................................................6 SOCIO-DEMOGRAPHIC CHARACTERISTICS OF THE RESPONDENTS ................................................10 Research Results .....................................................................................................................................................12 I ACCESS TO AND USE OF MEDIA AND ARTIFICIAL INTELLIGENCE ................................................12 II CRITICAL READING ...................................................................................................................................14 MEDIA IMPACT ON RESPONDENTS .......................................................................................................15 CONTENT VERIFICATION.........................................................................................................................16 SOURCE CREDIBILITY ..............................................................................................................................17 THE ROLE OF INSTITUTIONS ...................................................................................................................19 III PARTICIPATION .........................................................................................................................................21 DIGITAL SKILLS .........................................................................................................................................21 ACTIVITIES ..................................................................................................................................................22 Conclusion and Recommendations ........................................................................................................................23

2 Research on media literacy of the general population in Serbia was carried out within the Media Literacy Index project, a requirement set out in the Strategy for the Development of the Public Information System in the Republic of Serbia for the 2020–2025 Period (hereinafter: the Media Strategy). The survey was conducted by telephone between 15 May and 15 June 2025 on a two-stage stratified representative sample of 800 respondents. The findings will be integrated in the formulation of a Media Literacy Index, based on the EAVI model1 that clusters media literacy into two main dimensions: environmental factors and individual competencies. The research explores personal skills and abilities needed to engage critically and actively with media, such as cognitive skills for understanding and analyzing media content, technical skills to use media tools, and participatory skills to communicate and contribute through media channels.

1 https://ec.europa.eu/assets/eac/culture/library/studies/literacy-criteria-report_en.pdf 3  Research highlights

The research shows that television, messaging app, and music-oriented radio stations continue to dominate the media landscape:

• 96.7% of respondents watch television, • 81.5% use messaging apps to communicate and exchange messages, and • 66.3% listen to the radio, mostly to music stations.

The research points to a lower use of social media. While this finding warrants further investigation, one possible explanation is that young people between 15 and 29 years of age – the most frequent users of social media – make up only about 16% of Serbia’s population2.

Social media are used by 57.5% of the respondents.

Research shows that high proportion of people who do not follow the news, either through the press or via online news portals. Print media are not read by as many as 77.4% of the respondents. Online news portals are not followed by 43.5% of the respondents.

Between a quarter and a third of the respondents use streaming platforms, for watching films and series (e.g. Netflix, Max) or for listening to audio content - music or podcasts (e.g. Spotify, Deezer). Music streaming platforms are used by 26.7% of the respondents. Video streaming platforms are used by 35.7% of the respondents.

Artificial intelligence (AI) platforms (e.g. ChatGPT) are used by a quarter of the respondents. AI platforms are used by 25.4% of the respondents. Among them: 67.8% use AI to check their own content. 30% never check AI-generated content before sharing it. 30% never edit or adapt AI-generated content before sharing it.

A large share of the respondents said they did not recognise harmful content in the media. 27.6% of the respondents rarely or never recognise fake news. 51.5% of the respondents rarely or never recognise conspiracy theories.

2 Population by Age and Sex, 2023, Statistical Office of the Republic of Serbia. https://data.stat.gov.rs/Home/Result/3104020201?languageCode=en-US 4  25% of the respondents rarely or never recognise hate speech. 23.2% of the respondents rarely or never recognise the promotion of someone’s personal interests. 71.9% of the respondents rarely or never recognise AI-generated content.

Consistent with these findings, the media were observed to exert a strong impact on the respondents. A strong impact is shown on one side as strengthening the views users already have, and on the other, as exerting a strong emotional impact.

59.5% said they rarely or never learned anything new in the media that changed their views on a political or social issue. 68.6% reported feeling overwhelmed by media information and content either frequently or on a daily basis. 58.3% said they were frequently or daily exposed to media content that provoked strong negative emotions.

When it comes to fact-checking, 38.4% of the respondents said they often, and 33.6% occasionally, believed information from the media that later turned out to be false, while almost half (47.9%) rarely or never verified information across multiple sources. Similarly, 65.8% rarely or never read comments below posts and as many as 90.3% rarely or never checked information on a fact-checking website. Respondents most often verified the truthfulness of media information by consulting someone they knew – a practice reported by as many as 89.7%. The most trusted sources of information, according to 68.5% of respondents, are friends, relatives and colleagues, while far fewer place their trust in the media. Television was rated as a mostly or fully reliable source of information by 36.9% of respondents; it was followed by messaging apps (32.4%), radio (22%), AI (20.5%), social media (17.6%) and online news portals (11.7%). The least trusted medium was the press, viewed as reliable by only 3.7% of respondents. Among the digital skills measured, just over half of respondents reported knowing how to disable location tracking (55.7%) or digital activity tracking (52.2%), and how to block or remove someone from their friends list (54.9%). The least developed skills were those relating to protection from harmful content (lacking in 86.2%), closely followed by the ability to protect against online fraud (86%), to report hate speech (83.9%) and to disable AI memory (82.5%). The research data show that most of the respondents use the internet and media passively – primarily to obtain information on important issues (94.9%) and to connect with others (65.5%). They engage in other activities to a far lesser extent.

5  About the Research

Research on media literacy of the general population in Serbia was carried out within the Media Literacy Index project, a requirement set out in the Strategy for the Development of the Public Information System in the Republic of Serbia for the 2020–2025 Period (hereinafter: the Media Strategy). This project is being carried out in cooperation with and with the support of the OSCE Mission to Serbia, with the aim of implementing the Action Plan for the Implementation of the Media Strategy, which highlights (in Section 5.1) the lack of continuous systemic and systematic measurement of citizens’ media literacy, i.e. the cumulative effects of different indicators significant for its development. Since the development of a tool for researching the media literacy of the general population constitutes the first step towards achieving the Media Strategy Action Plan3, an initial working version of the media literacy index was first prepared in order to identify the most important indicators and the available data relating to them. Where it was impossible to obtain data concerning the key dimensions of media literacy, these were included as a battery of questions in the questionnaire. In accordance with the EAVI model4, the basic dimensions of media literacy covered by this questionnaire relate to individual competences (access and use, i.e. the critical dimension) and social competences (communication and participation). Emphasis was placed on the critical dimension of media literacy.

Drawing on the EAVI media literacy model and its core user-related dimensions, the following research questions were formulated:

RQ 1. How do Serbia’s citizens use (traditional and digital) media and artificial intelligence? RQ 2. What critical skills do they develop in relation to the media and artificial intelligence? RQ 3. What forms of communication and participation can we identify in the digital space and which digital skills are the citizens developing?

Sample The sample for the research on media literacy among the general population of the Republic of Serbia was based on the probability theory. Random selection is a key assumption for producing unbiased estimates of population parameters, as randomly chosen units are expected to reflect the overall characteristics of the population. The sample was defined as a two-stage stratified representative sample of Serbia’s population (since reducing error in estimated parameters is achieved by minimising variability

3 Activity 5.1.1. of the 2020-2025 Media Strategy provides for the establishment of a media literacy index in the Republic of Serbia in accordance with European methodology. The Media Strategy is available in Serbian at: https://www.media.srbija.gov.rs/medsrp/dokumenti/medijska_strategija210_cyr.pdf. 4 https://ec.europa.eu/assets/eac/culture/library/studies/literacy-criteria-report_en.pdf. 6 between the units studied), comprising 800 respondents. This division process is known as stratification, with the resulting segments referred to as strata. The basic strata were the main territorial units: Belgrade, Vojvodina and Central Serbia. Within each of these, urban and rural substrata were defined, giving a total of six strata. The sample design enables extrapolation to the population as a whole, meaning that the data relate to the population of Serbia (the territory of Kosovo and Metohija was not included in the research), between 18 and 75 years of age. The results can also be disaggregated by sociodemographic characteristics – sex (men and women), age (18–29, 30–44, 45–59 and 60–75), education (incomplete and completed primary education, secondary education, and higher or university education), and region (Belgrade, Vojvodina and Central Serbia). This sample makes it possible to draw conclusions at the national level, with a margin of error of 3.5 percentage points.

Questionnaire The semi-standardised questionnaire on the media literacy of the general population contains closed questions, with the exception of one open question and identification characteristics of the respondents (sex, age and educational attainment). The only open question was the one asking the respondents to specify their year of birth. Closed questions contained pre-formulated response options.

When it designed the questionnaire, and to ensure comparability, the research team drew on media literacy studies of the general population in European countries (Ofcom, 2024; Oxford, 2019) and in neighbouring countries (Paško, 2023; Raičević & Vranjanac, 2023), as well as on data from the Statistical Office of the Republic of Serbia (SORS), which had included a number of media literacy questions in its annual surveys on ICT use until 2024. In addition to e-skills, these included four questions concerning the respondents’ experience with disinformation – whether they had encountered it, whether they had checked it, how they had checked it and, if not, why not – as well as a set of questions on privacy (Kovačević et al., 2023). In 2024, SORS replaced groups of questions relating to digital skills and privacy protection by sets of questions relating to the industry – use of smart devices and green ICT (Josipović et al., 2024). This means that there is no longer any institution in Serbia that measures even the minimum of these important competences among citizens on a representative sample and on an annual basis.

Most media literacy research studies begin with a set of questions on access to and use of different media. These questions relate to the frequency and dynamics of media, device and internet use, as well as the types of media and media content, whilst focusing on news media and social media. The second element of media literacy repeatedly examined in various studies concerns critical thinking. This mainly consists of a battery of questions on trust in the media, often focusing on the area of information and disinformation. Participation is the third element of media literacy in the analysed studies, which

7 encompasses communication and creation, as well as civic engagement. The Oxford Internet Institute questionnaire (OXI, 2019) includes detailed batteries of questions on content production and a wide range of other online activities. SORS has placed emphasis on the economic aspect, i.e. on skills related to safe online shopping and communication with the public administration. Croatian research also emphasised content production skills (Paško, 2023), while the Montenegrin study linked this part to its key research topic - disinformation (Raičević & Vranjanac, 2023).

The questionnaire developed within this research places emphasis on the dimensions of media literacy relating to critical thinking and media and content evaluation, wherefore all other questions (on e.g. media use) were adapted accordingly. The researchers examined ways of use of various media, both traditional and digital, including artificial intelligence, which is becoming increasingly important in generating, creating and editing content, as well as in fact-checking. Activities covering digital skills, content creation, verification and dissemination, as well as the role of artificial intelligence, were also in focus. Although media literacy includes popular formats, they were covered to a lesser extent due to the constraints of space and the efficiency required by the CATI method. The questionnaire primarily examined news media, artificial intelligence, and the challenges they pose to the development of media literacy.

The research was conducted between 15 May and 15 June 2025 across Serbia. The telephone survey technique was used, with the assistance of the Computer Assisted Telephone Interview (CATI) platform, which provides direct computer support to different phases of the research through specialised software. The SPSS statistical analysis programme was used to process the data collected through the questionnaire.

The data obtained through the research reflect the respondents’ subjective views and opinions. With regard to critical thinking, this type of analysis needs to be complemented by qualitative methods. Coverage of all the important elements and questions by a single questionnaire is definitely not only a complex task requiring considerable resources but may also not always be feasible. Researchers involved in the evaluation of the European model for analysing media literacy indicators (EC, 2011) do not recommend this. “Given the breadth of contexts and behaviours associated with media literacy, a simple 20-minute survey, no matter how well-designed, cannot provide the comprehensive measures necessary to inform policy […] Attempting a comprehensive measure within a single survey is neither feasible, nor recommended.” (Bulger, 2012, p. 95). They propose a so-called modular approach, consisting of two parts. One study would focus on the basic issues of media literacy, which are quantified regularly, while

8 the other would be concerned with developing a rotating segment of the research, focusing “on specific components of media literacy and could be flexible to adapt to new findings or priorities.” (ibid.).

References

Bilić, Paško. 2023. Medijska pismenost građana Republike Hrvatske. istraživački izvještaj. Institute for Development and International Relations (IRMO). https://gong.hr/wp-content/uploads/2023/07/MOV- Medijska-pismenost-gradana-Republike-Hrvatske.pdf. Bulger, Monica. 2012. Measuring media literacy in a national context: challenges of definition, method and implementation. Media Studies, 3(6), 83-104. https://hrcak.srce.hr/file/141857. Celot, Paolo & Perez Tornero, Jose Manuel (2009) Study of the Assessment Criteria from Media Literacy: A comperhensive view of the concept of media literacy and understanding how media literacy levels in Europe should be assessed. For the European Commission Directorate General Information Society and Media; Media Literacy Unit. Brussels: European Association for Viewers’ Interests. Приступљено 17.9.2018. на URL адреси: http://ec.europa.eu/assets/eac/culture/library/studies/literacy-criteria-report_en.pdf. European Commission. 2011. Testing and Refining Criteria to Assess Media Literacy Levels in Europe Final Report. Commissioned by the European Commission Directorate-General for Information Society and Media; Media Literacy Unit. Danish Technological Institute & European Association for Viewers’ Interests. https://eavi.eu/wp- content/uploads/2017/08/study_testing_and_refining_ml_levels_in_europe.pdf- Josipović, Branko, Šutić, Vladimir, Rajčević, Uroš and Minaeva, Ivana. 2024. Usage of Information and Communication Technologies in the Republic of Serbia, 2024. Belgrade: SORS. https://publikacije.stat.gov.rs/G2024/PdfE/G202416019.pdf. Kovačević, Miladin, Šutić, Vladimir, Rajčević, Uroš and Minaeva, Ivana. 2023. Usage of Information and Communication Technologies in the Republic of Serbia, 2023. Belgrade: SORS. https://publikacije.stat.gov.rs/G2023/PdfE/G202316018.pdf. Ofcom. 2024.Adults’ Media Use and Attitudes Report. https://www.ofcom.org.uk/media-use-and- attitudes/media-habits-adults/adults-media-use-and-attitudes-2024-interactive-report. Oxford Internet Institute – OXI. 2019. Oxford Internet Survey 2019 General Questionnaire. https://oxis.oii.ox.ac.uk/wp-content/uploads/sites/16/2019/11/OxIS-Questionnaire-2019.pdf Press Council. 2020. Leksikon medijske pismenosti. https://savetzastampu.rs/wp-content/uploads/2020/12/3- LEKSIKON-MEDIJSKE-PISMENOSTI.pdf. Raičević, Vladimir & Vranjanac, Marija. 2023. Media Literacy in Montenegro. Public opinion research. Ipsos. https://www.osce.org/files/f/documents/0/a/564889.pdf. Strategijа razvoja sistema javnog informisanja u Republici Srbiji za period 2020–2025. godina. „Službeni glasnik RS”, broj 30/18. https://media.srbija.gov.rs/medsrp/dokumenti/medijska_strategija210_cyr.pdf

9 SOCIO-DEMOGRAPHIC CHARACTERISTICS OF THE RESPONDENTS

Slightly more than half of the respondents were women, reflecting the gender structure of the population of the Republic of Serbia. The sample comprised 409 women (51.1%) and 391 men (48.9%). As per the age breakdown of the respondents, 31.9% were aged 60 and over, almost a quarter (24.3%) were between 45 and 59 years of age, while 27.6% were between 30 and 44 years of age. Those aged 18–29 accounted for the smallest number of respondents - 16.2%.

Table 1: Respondents by Sex (n and %): Table 2: Respondents by Age (n and %):

Sex n % Age Group n % Men 391 48.9 18 - 29 130 16.2 Women 409 51.1 30 - 44 221 27.6 Total: 800 100 45 - 59 194 24.3 60 - 75 255 31.9 Total: 800 100

In terms of educational attainment, more than half (55.9%) of the respondents had completed secondary education, a fifth (20.5%) had a tertiary qualification, 6.5% had some higher education but had not earned a degree, while 17.1% had only incomplete or complete primary education (Table 3).

Table 3: Respondents by Education Level (n and %)

Highest Educational Attainment n % Incomplete and Complete Primary 137 17.1 Education Secondary Education 448 55.9 Incomplete Tertiary Education 52 6.5 Tertiary Education 163 20.5 Total: 800 100

Most of the respondents were employed or retired (50.9% and 17.9% respectively). The unemployed accounted for 5.9% and college students for 6.3% of the respondents. Nine percent of the respondents were homemakers, while 10% were engaged in agriculture (Table 4).

Table 4: Respondents by Employment Status (n and %)

Employment Status n % Employed 408 50.9 Unemployed 47 5.9 Retired 143 17.9 Student 50 6.3 Agricultural worker 80 10 Homemaker 72 9 Total: 800 100

10 Most of the respondents – nearly a half of them (48.3%) – were living in Central Serbia, while 26.3% lived in Vojvodina and 25.4% lived in Belgrade. More than half of the respondents (58.9%) lived in urban settlements, while 41.1% lived in rural areas.

Table 5: Respondents by Region (n and %) Table 6: Respondents by Settlement Type (n and %):

Region n % Central Serbia 386 48.3 Settlement Type n % Belgrade 204 25.4 Urban 471 58.9 Vojvodina 210 26.3 Rural 329 41.1 Total: 800 100 Total: 800 100

Most respondents (39.8%) lived in households with a monthly income ranging from 60,000 RSD to 150,000 RSD5. A substantial share – more than a fifth, specifically 22.4% – lived in households with a monthly income not exceeding 60,000 RSD. Monthly household income of between 150,000 RSD and 200,000 RSD was reported by 18.3% of respondents, while only 8% stated that their household income exceeded 200,000 RSD. Slightly over a tenth (11.5%) of the respondents did not wish to disclose their household’s monthly income.

Table 6: Respondents by Average Household Income (n and %)

Average Monthly Household Income n % Under 60,000 RSD 179 22.4 Between 60,000 and 150,000 RSD 319 39.8 Between 150,000 and 200,000 RSD 146 18.3 Over 200,000 RSD 64 8 Prefer not to answer 92 11.5 Total: 800 100

5 The average net salary in Serbia in May 2025 was 107,705 RSD, while median net income was 84,408 RSD: https://stat.gov.rs/en-us/vesti/statisticalrelease/?p=17056 11  Research Results

I ACCESS TO AND USE OF MEDIA AND ARTIFICIAL INTELLIGENCE

This section of the research measured the frequency with which specific media and artificial intelligence platforms were used. The first set of questions in the questionnaire asked respondents to assess how often they used particular media. Respondents answered by selecting one of five options on a five-point scale: 1 – Never, 2 – Rarely, 3 – Occasionally, 4 – Often, and 5 – Daily (see Figure 1).

Figure 1: Frequency of (Traditional and Digital) Media and Artificial Intelligence Platform Use (%)

90 80 70 60 50 40 30 20 10 00 Streaming Music AI Platforms Online news platforms Messaging TV Radio Press streaming Social media (e.g. portals (films and Apps platforms ChatGPT) series) Never 03 34 77 44 64 73 43 19 75 Rarely 14 25 14 17 13 06 09 05 06 Occasionally 27 27 06 25 17 15 19 20 13 Often 21 10 01 12 05 05 17 26 05 Daily 35 04 01 03 01 01 12 30 01

According to the data obtained in this research, television is the dominant media outlet in terms of frequency of use – more than half of the respondents watch it either often (20.9%) or every day (35.3%). Only 3.3% of the respondents said they never watched television. Communication and messaging applications (such as Viber, WhatsApp, etc.) are used almost as much – 29.8% of respondents use them daily, 26.4% often, around a quarter use them rarely or occasionally (5.3% and 20% respectively), while about a fifth (18.5%) never use these applications.

It is also clear that traditional media, such as the press and radio, have lost substantial ground to television. A third of the respondents (33.7%) said they never listened to the radio, a quarter (25.1%) reported listening to it only rarely, while only 3.6% said they listened to it every day. The fewest respondents read the press: 77.4% of the respondents reported never reading it, while a further 14% said they read it only rarely. This leaves slightly over 2.5% who read newspapers often or every day.

12 Over half of the respondents keep informed through online news portals, while 43.5% never use them. Sixteen percent read them rarely and a quarter (24.6%) occasionally. A further 11.9% use portals often and 3.1% every day - figures almost identical to those of the most regular radio listeners.

Streaming platforms are not widely used – 73.3% of the respondents do not use music streaming services (e.g. Spotify, Deezer), while 64.3% do not use such services for watching films and series (e.g. Netflix, Max, Amazon Prime, Disney+). Only 1% use them every day, around 5% use them often, while most of those who engage with such content do so occasionally – 17.1% of respondents watch and 15% listen to content via dedicated streaming services.

Social media are used daily by 12.4% and often by 16.8% of the respondents. One in five (19.2%) use them occasionally, 9.1% rarely, while 42.5% do not use them at all. It is important to note that these figures refer solely to the frequency of use over the past month, not to the number of respondents with social media accounts.

Interestingly, similar shares of respondents use artificial intelligence (AI) and streaming platforms. Most respondents have never used AI (74.6%), 5.9% use it rarely, 12.9% occasionally and only 5.3% and 1.3% use it often or daily, respectively. Given that both streaming services and AI tools have only recently become available, it will be interesting to see how quickly their use becomes widespread.

USE OF ARTIFICIAL INTELLIGENCE Respondents were also asked a specific set of questions (Figure 2) aimed at measuring diversity of AI use. Three-quarters of them (around 75%) reported that they did not use AI, while 25.4% stated that they did. Respondents who used AI mainly did so to check and refine content they had already created (67.8%) and for researching and generating ideas before creating content (56.4%). Fewer than half reported using AI to obtain guidance on designing and structuring content (46.2%) or for content creation itself (45.7%), while the smallest proportion used it to develop content placement and promotion strategies.

13 Figure 2: Reasons for Using Artificial Intelligence among Respondents Who Reported Using It (%)

67.8 61.6 54.3 56.4 53.8 45.7 43.6 46.2 38.4 32.2

Create content (text, Research and generate Obtain guidance on Develop content Check and improve photograph, video) ideas before creating designing and structuring placement and content I previously content content promotion strategies created

Yes No

Respondents were also asked how often they adapted and checked AI-generated content before sharing it (Figure 3). About one-third (32%) reported that they always adapted such content (32%) and almost as many (30.1%) that they never did so before sharing it; 28.2% said they always checked the content and 30% reported that they never did so before sharing it. Roughly one in five respondents reported carrying out both activities often, while around one in ten did so rarely.

Figure 3: Frequency of Checking or Adapting AI-Generated Content (%) When you use AI, how often do you:

30.1 32 30 28.2

18.7 20 12.7 10.1 9.1 9.1

ADAPT IT before sharing it CHECK IT before sharing it

1 - Never 2 - Rarely 3 - Occasionally 4 - Often 5 -Always

II CRITICAL READING

This part of the questionnaire focused on assessing the respondents’ self-perceived ability to recognise various types of media content, whether or not it involved some form of AI-generated disinformation or manipulation. The respondents were given the option of answering on a five-point scale: 1 – Never, 2 – Rarely, 3 – Occasionally, 4 – Often and 5 – Always (Figure 4).

The respondents expressed the greatest confidence in their ability to identify hate speech and promotion of personal interests in the media: 7% of them stated that they could always recognise such content with certainty, while more than 40% reported that this was often the case – slightly more so in relation to the 14 promotion of personal interests (43.9%) than hate speech (40.4%). They were less confident about their ability to recognise fake news and conspiracy theories – only a very small share could confirm with certainty that they were always able to identify such content (1.4% for fake news and 0.6% for conspiracy theories). However, more than a third of the respondents reported that they recognised fake news with certainty often (33%) or occasionally (38%). By contrast, far fewer felt confident about recognising conspiracy theories in the media – 13% said they were often certain they could identify them, while 34.9% said they occasionally did so with certainty. The respondents were the least confident in their ability to identify AI-generated content – as many as 71.9% stated that they were never sure they recognised it.

Figure 4: Content Recognition (%)- How often are you sure you recognised the following in the media?

80 71.9 70 60 50 40.4 43.9 38 34.9 40 33 31.9 27.6 25.8 30 17.8 19.6 14.4 13.6 20 9.8 13 10.6 7 9.6 7.1 10.413.6 10 1.4 0.6 3.7 0.4 0 Fake or deliberately Conspiracy theories Hate speech Promotion of personal AI-generated content misleading news interests

1 - Never 2 - Rarely 3 - Occasionally 4 - Often 5 - Always

MEDIA IMPACT ON RESPONDENTS The questionnaire also addressed the media’s impact on the users’ views and feelings (Figure 5). When asked how often they had learned something in the media that changed their views on a political or social issue, more than half of respondents answered that this happened rarely (32.1%) or never (27.4%). A further 27.9% said their opinions were swayed occasionally, while 11.5% reported that the media often had such an impact on them.

The media have a strong impact on the respondents’ emotional responses. Most reported feeling overwhelmed by emotions and media content either often (41.7%) or every day (26.9%). Only 2.6% said they had never experienced this, while 8.5% reported it happening rarely. A fifth of the respondents (19.9%) said they occasionally felt overwhelmed by emotions triggered by media content.

More than half of the respondents also said that media content often (47.9%) or daily (10.4%) triggered strong negative emotions such as sadness, anger, shame, disgust, fear, etc. A substantial share (31.8%)

15 reported feeling this way occasionally, while only 2.4% said they had never experienced such feelings because of something they had seen or heard in the media.

Figure 5: Media Impact on Users (%) – How often have you:

60

50 47.9 41.7 40 32.1 31.8 30 27.4 27.9 26.9 19.9 20 11.5 10.4 8.5 7 10 2.6 2.4 0.3 0.8 0.4 0.5 0 Learned something NEW from the media that Felt OVERWHELMED BY INFORMATION and content Been exposed to information from the media that CHANGED your views on a political or social issue? from the media? provoked STRONG NEGATIVE EMOTIONS?

1 - Never 2 - Rarely 3 - Occasionally 4 - Often 5 - Always Don’t know

CONTENT VERIFICATION Questions relating to the practice of verifying information, i.e. media content, were also scaled from 1 to 5 (1 – Never, 2 – Rarely, 3 – Occasionally, 4 – Often, 5 – Always). The respondents answered questions about ways of checking suspicious information they encountered in the media – from cross-checking across different media sources, reading comments and consulting people they know, to using fact- checking resources or artificial intelligence (Figure 6).

Respondents were least likely to verify information through fact-checking websites or applications: 84% said they never did so, 6.3% rarely, 5.8% occasionally, while only 3.1% reported doing so often and just 0.8% daily. Likewise, only a small proportion used artificial intelligence for this purpose. Just 0.3% said they used it daily, 6.1% often, 11% occasionally, 6.8% rarely, while 75.5% reported never having used AI to verify suspicious information in the media.

When respondents found a piece of information from the media suspicious, they most often verified it by consulting people they knew or by checking other media sources, with acquaintances and close contacts having a slight edge over the media. Only 10.3% said they had never sought verification from people they knew, while 32.6% said they did so rarely. Some 35.5% of the respondents reported verifying information in this way occasionally, 18.8% often and 2.8% daily. By comparison, one in five respondents turned to other media sources, often (18.5%) or daily (1.8%). A further 31.5% did so

16 occasionally and 26% rarely. Over a fifth (21.9%) never checked suspicious information against multiple sources.

Half of the respondents reported never reading comments on suspicious content or information, while only 1.4% said they did so daily. Among those who read comments, a fifth (20.5%) did so occasionally, 15.8% rarely and 12.3% reported doing so often.

An additional question - on personal experience with fake or misleading news - was added to questions about verifying suspicious content and information published by the media. Asked how often they had believed information from the media which later turned out to be false, 38.4% of respondents said this often happened to them and 33.6% that it happened occasionally. A fifth of the respondents (20.8%) reported it had happened rarely, while 6.4% claimed it had never happened to them.

Figure 6: Respondents’ Personal Experience with Fake or Misleading News (%) – As regards suspicious content or information in the media, how often have you:

Believed Checked it on a Checked it with information in Checked it across Read the fact-checking Checked it by someone you the media that multiple sources? comments? website or using AI? know? later turned out application? to be untrue? 1 - Never 21.9 50 84 75.5 10.3 6.4 2 - Rarely 26 15.8 6.3 6.8 32.6 20.8 3 - Occasionally 31.5 20.5 5.8 11 35.5 33.6 4 - Often 18.5 12.3 3.1 6.1 18.8 38.4 5 - Daily 1.8 1.4 0.8 0.3 2.8 0.4 Don’t know 0.3 0 0 0.3 0 0.4

1 - Never 2 - Rarely 3 - Occasionally 4 - Often 5 - Daily Don’t know

SOURCE CREDIBILITY Questions concerning the reliability of sources of information aimed to measure whom the respondents trust the most – the media, artificial intelligence, or family, friends and colleagues. The respondents selected answers on a scale from 1 to 5, where 1 indicated a source that was not reliable at all and 5 a source that was totally reliable. They were also given the option of stating that they were unable to assess 17 the reliability of a source. (Figure 7). The respondents place the greatest trust in people they know - friends, colleagues and family. More than 68% consider them either fully (33.1%) or mostly reliable (35.4%) sources of information, while 22.9% consider them both reliable and unreliable.

Figure 7: Reliability of Sources of Information (in %)

1 -Totally unreliable 2 - Mostly unreliable 3 - Both reliable and unreliable 4 - Mostly reliable 5 - Totally reliable Don’t know/Can’t assess

77.9

74.6 59.6

50.4

43.5

35.4 34.4

33.1 32.3 27.5

26.6 23.8 23.3

22.9 22.5 20.3

14.9 14.5 14.5

14.1 14.1 13.6

10.9 10.1 9.4 7.5

7.3 6.7

6.1 5.8

5.6 4.9

4.4

4.1 3.5

3.5 3.3

2.6

2.4 2.3 1.6 1.1 0.6

0.6

0.4 0.4

0.1 1

Television Radio Press News Portals Social Media Messaging Artificial Family, Apps Intelligence Friends, Colleagues

Given that many respondents said they were unable to assess the reliability of some types of media, their responses may be linked to how frequently they use them: respondents tend to feel more confident judging the reliability of media they use often and understand well, and vice versa. For example, 74.6% of the respondents could not assess how reliable artificial intelligence is. Interestingly, a similar pattern emerged with the press, where 77.6% were unable to judge its credibility as a source of information. Likewise, when asked about radio, 59.6% reported being unable to assess its reliability. By contrast, only 3.5% of respondents could not judge the reliability of television. Among those who did evaluate media credibility, most were ambivalent, viewing them as both reliable and unreliable — a finding likely related also to the general nature of the question and the diversity of outlets within each category.

Of the 96.5% of the respondents who expressed an opinion on the reliability of television, 36.9% considered it a fully or mostly reliable, compared with 25.2% who qualified it as an unreliable or mostly unreliable source of information, while 34.4% regarded it as both reliable and unreliable. Around 40% of the respondents expressed an opinion about radio, with 14.5% considering it both reliable and

18 unreliable. A much greater share of respondents, however, regarded it as reliable (22% in total) compared with those who considered it unreliable (3.9%).

Most respondents who evaluated news portals were ambivalent (27.5%). Some 11.7% considered them reliable sources, compared with 10.4% who distrusted them, while more than half were unable to assess their credibility as sources of information. The fewest respondents expressed an opinion about the reliability of the press — only 22.1%. However, the answers of just these respondents show that the press is rated as the least credible source: 11.7% qualified it as unreliable or mostly unreliable, compared with just 3.7% who consider it a credible source.

The figures are almost evenly divided between those who trust (17.6%) and those who distrust (16.4%) social media as sources of information. However, more respondents considered them both reliable and unreliable (22.5%) or were unable to assess their credibility (43.5%). Messaging applications and artificial intelligence were rated as more credible than social media. However, a much larger share of respondents expressed an opinion about the reliability of messaging apps – more than three-quarters (76.2%), which corresponds to the proportion of respondents who reported using them every day, often, or at least occasionally. Of them, 32.3% considered messaging apps both reliable and unreliable, 11.5% unreliable and 32.4% reliable or mostly reliable sources of information. Although fewer respondents evaluated artificial intelligence (though more than the press!), the smallest proportion were ambivalent (4.1%) about the credibility of this source and only 0.8% regarded it as unreliable. A fifth (20.5%) of the respondents assessed artificial intelligence as a credible source of information.

THE ROLE OF INSTITUTIONS An important aspect of media and information literacy concerns attitudes towards the role of the electronic media regulatory authority (REM), as well as understanding the difference between public service broadcasters and commercial media. Respondents were asked to express their (dis)agreement with the four statements they were given on a scale of 1 to 5, where 1 indicated complete disagreement and 5 complete agreement. The first two statements regarded REM and its remit and whether public complaints to REM about unprofessional media conduct were a worthwhile course of action (Figure 8).

19 Figure 8: REM’s Role (%)

81.4

48.2

18.4 8.9 7.8 12 12.9 7.1 0.8 0.1 1 1.4

REM is tasked with monitoring all media and for warning or Public complaints of unprofessional media practices or harmful penalising outlets that broadcast inappropriate content, and is to content to REM are a worthwhile course of action. carry out these duties free from political or economic influence.

1 – Totally disagree 2 – Mostly disagree 3 – Both agree and disagree 4 – Mostly agree 5 – Totally agree Don’t know

Almost all respondents agreed (8.9% mostly and 81.4% completely) with the statement that REM was charged with monitoring the work of all media, issuing warnings and imposing sanctions against outlets broadcasting inappropriate content, and that it was to carry out these duties free from political or economic influence. Only 1% of the respondents disagreed with this statement or were ambivalent. However, the answers were somewhat different when it came to public complaints about the outlets’ unprofessional conduct. Around 60% agreed (12.9% mostly and 48.2% completely), while 12% were undecided about whether public complaints to REM about unprofessional media conduct or harmful content were a worthwhile course of action. Only 8.5% mostly or completely disagreed with this statement.

The other two statements regarded the responsibility and obligations of the media, both commercial and public service broadcasters (Figure 9).

Figure 9: Responsibility of Commercial Media and Public Service Broadcasters (%)

67.8

52

14.3 13.8 12.6 6.1 8.4 8.1 5.4 4.4 5 2.1

Commercial media are exempt from social responsibility because Public service broadcasters should bear a higher level of they are privately owned. responsibility than private media.

1 – Totally disagree 2 – Mostly disagree 3 – Both agree and disagree 4 – Mostly agree 5 – Totally agree 77 Don’t know

20 More than 80% of the respondents opined that public service broadcasters should bear a higher level of responsibility than private outlets (12.6% mostly agreed and 67.8% completely agreed with the statement). Only 6.5% disagreed, while 8.1% were ambivalent. Most respondents expressed complete (52%) or partial disagreement (14.3%), while 13.8% were ambivalent about whether commercial media were exempt from social responsibility because they were privately owned. Slightly over a tenth of the respondents (11.5%) agreed (mostly or completely) with the statement that commercial media need not be socially responsible because they were privately owned.

III PARTICIPATION

This part of the research focused on the respondents’ self-assessment of their digital skills, as well as on their use of media for various activities, such as connecting with other users sharing their interests and participation in social life.

DIGITAL SKILLS The part of the research dedicated to digital skills provides particular insight into the development of various types of knowledge and experience in using digital information and communication technologies, such as the internet, mobile phones, computers and various software applications designed for communication and messaging, social media, etc., with focus on communication, participation, content creation and protection against harmful content and interactions. The respondents were asked a series of yes/no questions concerning practices related to privacy protection and safeguarding against harmful content (Figure 10).

According to the data obtained, more than half of the respondents know how to turn off location tracking (55.7% answered “yes” to this question, compared with 44.3% who said ‘no’), or how to block or remove someone from their friends list (54.9% knew how, while 45.1% did not). Similarly, just over half of respondents (52.2%) know how to disable tracking of their digital activity, such as their browsing history.

The lack of skills is more evident in certain other practices relating personal data protection and protection against harmful content. For example, 48.4% of internet users said they could tell when they could share their personal data, while more than half (51.6%) said they could not. A much smaller share – only 14% – claimed they knew how to protect themselves against online fraud (identity theft, spam, phishing, financial scams, etc.). Only 17.5% of the respondents knew how to disable AI memory. With regard to protection from harmful content, only 13.8% of the respondents said they knew how to protect themselves from material such as explicit violence, hate speech, pornography, insults, humiliation and 21 the like. Respondents gave similar answers in response to the question whether they knew how to report hate speech on social media – 16.1% answered in the affirmative.

Figure 10: Digital Skills (%)

Do you know how to block or remove someone from your friends list? 54.9 45.1 Do you know how to protect yourself from online fraud (identity theft, spam, phishing, financial scams, etc.)? 14 86 Do you know how to protect yourself from harmful content (explicit violence, hate speech, pornography, insults, humiliation, etc.)? 13.8 86.2

Can you tell when it is safe to share your personal data? 48.4 51.6

Do you know how to turn off the memory function in AI tools? 17.5 82.5 Do you know how to turn off tracking of your online activity (browsing history)? 52.2 47.8

Do you know how to turn off location tracking? 55.7 44.3

Do you know how to report hate speech on social media? 16.1 83.9

1 – Yes 2 – No/Unsure

ACTIVITIES The research data indicate that the internet is used primarily for obtaining information and only sporadically for exchanging opinions and connecting with other people who share the same interests or concerns (Figure 11).

Figure 11: Most important online activities (%)

86.6 81.7

66.3

51.6

34.5 29.631.1 28.3 20.6 23.9 22.720.8 13.1 16.114.3 12 9.1 7.3 5.1 4.8 3.3 6.4 5.9 0.5 0.9 0.4 0.1 1.9 1.1

Inform myself about Connect with people with Freely express my opinion Advocate a specific idea Take to task public figures Participate in or organise important social, political similar interests or about important issues or goals events, protests or and other issues concerns campaigns

1- Never 2 - Rarely 3 - Occasionally 4 - Often 5 - Daily Don’t know

More than half of the respondents use the internet to inform themselves about important social, political and other issues (31.1% often and 20.6% daily, around 30% occasionally), while only 5% use it regularly

22 to express their own views or opinions. Some 28.3% of respondents occasionally connect with others, while more than half do so rarely or never (altogether 58.4%). Almost three-quarters of the respondents are not in the habit of expressing their opinions freely in the online world, with more than half (51.6%) stating that they never do so.

A negligible number of respondents see the internet as a platform for advocating ideas and goals or for organising broader social activities, protests or campaigns. Only 14.3% engage occasionally in advocacy (3.3% do so often), while 16.1% do so rarely. Some 66.3% said they had never used the media or internet for such purposes. As per participation in or organisation of events, as many as 86.6% of the respondents said they had never used the internet or media for that purpose, 6.4% said they did so rarely, 5.9% occasionally and only 1.1% often. The internet is also not used for holding public figures to account – 9.1% of the respondents said they rarely did so, while as many as 81.7% never did. Only 7.3% said they occasionally did so and just 1.9% reported that they did so often.

Conclusion and Recommendations

The research data clearly show that Serbia’s citizens predominantly watch television and, to a lesser extent, follow online news portals, whilst disregarding other traditional media – the radio and the press. Particularly noteworthy is the low level of trust the respondents expressed in print media and online portals. Approximately one quarter of all respondents use AI (around 25%), and they mostly use it to check their created content or research before content creation.

Research shows that two-thirds of respondents place their trust in their colleagues, friends, peers, and family and have less trust in media. A very small number of respondents use fact-checking resources and rely predominately on the people they know when verifying the truthfulness of media information. While they feel sure they can recognize content such as hate speech or personal promotion, they are not as certain when it comes to fake news and conspiracy theories—only a very small share could confirm with certainty that they were always able to identify such content (1.4% for fake news and 0.6% for conspiracy theories). This shows a lack of developed critical media reading skills, although most of them feel overwhelmed by media content. The biggest issue still is with AI, where the respondents are the least confident—71.9% stated that they were never sure if they could recognize AI-generated content.

The respondents use the internet passively, mostly for obtaining information and to connect with others via messaging and social networks. Only 5% of respondents use the internet regularly to express their

23 own views or opinions, over 60% of them have never used it as a platform for advocacy, and over 86% have never used it to participate in or organize an event.

The lack of skills is evident in practices relating to personal data protection and protection against harmful content. Just over half of respondents know how to disable location tracking (55.7%) or digital activity tracking (52.2%) and how to block or remove someone from their friends list (54.9%). The least developed skills were those relating to protection from harmful content (lacking in 86.2%), closely followed by the ability to protect against online fraud (86%), to report hate speech (83.9%), and to disable AI memory (82.5%).

Following these results, several recommendations can therefore be drawn:

1. Maintaining media pluralism is definitely a major challenge for media policy in Serbia today, not only in terms of media content, but also in terms of the platforms via which such content can be made available to the public. 2. The low level of trust in news media, particularly the press and online editions, points to the need for promoting high-quality investigative journalism and media that uphold journalistic professionalism, provide reliable information and create fertile ground for the development of critical awareness and critical thinking in the public sphere – preceded by ensuring the protection of media freedoms and freedom of expression. This is also linked to the need to establish effective systemic support for vulnerable and smaller media, especially local outlets, as opposed to the existing practice of co-financing media content through competitive calls. Several studies (see Ranković, 20216; Matović & Purić, 20187; Nikoletić, 20208; Savić & Milošević, 20219) have highlighted problems in the allocation of funds under these schemes, pointing to bias in the selection of media, vague criteria and non-transparency of the competition process. 3. The respondents recognise the important role of the media regulator and the need for it to monitor and supervise, and, where necessary, penalise outlets that broadcast inappropriate content

6 Ranković, Larisa. 2021. MEDIJSKI KONKURSI MINISTARSTVA KULTURE I JAVNOG INFORMISANJA U 2020. I MEDIJSKI PLURALIZAM. BIRN & OSF. Available in Serbian at: https://nuns.rs/media/2021/11/Medijski-konkursi- Ministarstva-kulture-i-javnog-informisanja.pdf. 7 Matović, Dragana and Purić, Biljana. 2018. Analiza obrazloženja komisija za dodelu sredstava na konkursima za sufinansiranje medijskih sadržaja: nejasno i netransparentno. BIRN. Available in Serbian at: https://safejournalists.net/wp- content/uploads/2019/08/Analiza-obrazlo%C5%BEenja-komisija-za-dodelu-sredstava.pdf. 8 Nikoletić, I. Danas, 28 May 2020. Za vlast istraživanje korupcije nije u interesu javnosti. Available in Serbian at: https://www.cenzolovka.rs/drzava-i-mediji/za-vlast-istrazivanje-korupcije-nije-u-interesu-javnosti/. 9 Vlajović Savić, Bojana and Milošević, Marija. 2021. PRATIMO MEDIJSKI KONKURS Analiza prvog konkursa za sufinansiranje medijskih sadržaja od javnog interesa u Kragujevcu. Res Publika. Available in Serbian at: https://www.gradjanske.org/wp-content/uploads/2022/01/Analiza-konkursa-2021-KG-bros%CC%8Cura.pdf. 24  or breach the law and ethical codes. In addition to the clear need for a regulator with the mechanisms, legitimacy and authority to carry out oversight and regulation responsibly and without bias, the findings also suggest that citizens themselves can be an important resource in safeguarding REM’s independence and accountable work. Empowering them to use various courses of action and to cooperate with the regulator would also encourage greater engagement in the media and digital spheres. 4. Given that a large share of respondents uses media and digital technologies passively and within private spheres, emphasis should be placed on programmes combining the strengthening of critical thinking with the development of digital skills. Such initiatives would enable citizens to engage with the media more actively and confidently - whether for protection or for civic engagement and participation in public life. These programmes should be designed and delivered to ensure accessibility and to meet the needs of various groups, including specific segments of the population - particularly those considered “vulnerable” and those who use certain types of media most intensively. Coordination among a range of stakeholders in the field of media and information literacy (the media and academic communities, civic, professional and expert associations, and governmental and non-governmental sectors) remains essential in this respect to facilitate the integration of diverse initiatives into a concerted effort to strengthen citizens’ capacities, especially in the area of critical media and information literacy. 5. Artificial intelligence presents new challenges for all respondents, as it calls for the application of the full range of media literacy competences - access, analysis and evaluation, as well as communication, creation and participation - and for the acquisition of new digital skills in line with the scope of emerging technologies. Monitoring trends and fostering cooperation across sectors (IT, education, media and government) are essential both for adapting legislation and regulatory frameworks and for providing training programmes that familiarise citizens with the new developments. Media literacy initiatives involving AI should be planned and designed to ensure they are accessible to various segments of the population.

25

Projekat “Jačanje slobode izražavanja i slobode medija u Srbiji“, Misija OEBS-a u Srbiji

MEDIJSKA PISMENOST GRAĐANA SRBIJE Izveštaj anketnog istraživanja Beograd, 2025.

Istraživački tim: rukovoditeljka istraživačkog projekta Marijana Matović, Fakultet političkih nauka – Univerzitet u Beogradu i glavne istraživačice Ana Milojević Fakultet političkih nauka – Univerzitet u Beogradu i Snežana Milin Perković; konsultant Aleksandar Tomašević – Institut za fiziku – Univerzitet u Beogradu i istraživačice Slavica Josifović i Milena Jović.

Ovaj projekat finansira Evropska unija

Napomena: Ovo istraživanje objavljeno je uz finansijsku pomoć Evropske unije. Stavovi izrečeni u ovoj publikaciji pripadaju isključivo autorima i njihovim saradnicima i ne predstavljaju nužno zvaničan stav Misije OEBS-a u Srbiji i nužno ne odražavaju stavove Evropske unije.

1 Sadržaj

Istaknuti nalazi istraživanja ..................................................................................................................................4 O istraživanju............................................................................................................................................................6 SOCIO-DEMOGRAFSKE KARAKTERISTIKE ISPITANIKA ......................................................................10 Rezultati istraživanja ..............................................................................................................................................12 I PRISTUP I KORIŠĆENJE MEDIJA I VEŠTAČKE INTELIGENCIJE .........................................................12 KORIŠĆENJE VEŠTAČKE INTELIGENCIJE ............................................................................................13 II KRITIČKO ČITANJE ....................................................................................................................................14 UTICAJ INFORMISANJA ............................................................................................................................15 PROVERA SADRŽAJA ................................................................................................................................16 KREDIBILITET IZVORA .............................................................................................................................17 ULOGA INSTITUCIJA .................................................................................................................................19 III PARTICIPACIJA ..........................................................................................................................................21 DIGITALNE VEŠTINE .................................................................................................................................21 Zaključak i preporuke .............................................................................................................................................23

2 Istraživanje medijske pismenosti opšte populacije u Srbiji sprovedeno je u okviru projekta Indeksa medijske pismenosti, kao deo zahteva definisanih u Strategiji razvoja sistema javnog informisanja u Republici Srbiji za period 2020–2025 (u daljem tekstu: Medijska strategija). Anketa je realizovana putem telefona u periodu od 15. maja do 15. juna 2025. godine na dvostepenom stratifikovanom reprezentativnom uzorku od 800 ispitanika. Rezultati istraživanja će biti uključeni u formulisanje Indeksa medijske pismenosti, zasnovanog na EAVI modelu1, koji medijsku pismenost deli na dve glavne dimenzije: faktore okruženja i individualne kompetencije. Istraživanje ispituje lične veštine i sposobnosti potrebne za kritičko i aktivno angažovanje sa medijima, kao što su kognitivne veštine za razumevanje i analizu medijskog sadržaja, tehničke veštine za korišćenje medijskih alata i participativne veštine za komunikaciju i doprinos kroz medijske kanale.

1 national-policies.eacea.europa20102025-Media-literacy-of-citizens-of-Serbia-research-report-Clean.docx

3  Istaknuti nalazi istraživanja

Istraživanje pokazuje da, među medijima, još uvek dominira televizija, 96,7% ispitanika gleda televiziju. u kombinaciji sa korišćenjem mesindžera u visokom procentu, 81,5% ispitanika koristi aplikacije za komuniciranje i razmenu poruka. i slušanjem radio stanica (koje pretežno emituju muziku). Radio sluša 66,3% ispitanika.

Istraživanje pokazuje trend manjeg korišćenja društvenih mreža. Iako je ovaj podatak potrebno dodatno istražiti, on se možda može objasniti time što u opštoj populaciji mladi uzrasta 15-29 godina, koji su najveći korisnici društvenih mreža, čine oko 16% ukupnog stanovništva2. Društvene mreže koristi 57,5% ispitanika.

Drugi trend koji je zabeležen istraživanjem je visok procenat onih koji ne prate vesti niti putem štampe i niti putem informativnih onlajn portala.

Štampu ne čita čak 77,4% ispitanika.

Informativne onlajn portale ne prati 43,5% ispitanika

Striming platforme koristi svaki treći, odnosno četvrti ispitanik, u zavisnosti od toga da li je reč o platformi za gledanje serija i filmova (npr. Netflix, Max) ili slušanje audio sadržaja - muzike ili podkasta (npr. Spotify, Deezer).

Muzičke striming platforme koristi 26,7% ispitanika.

Striming platforme za serije i filmove koristi 35,7%

Platforme veštačke inteligencije (npr. ChatGPT) koristi svaki četvrti ispitanik,

Platforme veštačke inteligencije koristi 25,4%.

2 Stanovništvo prema starosti i polu. 2023. Republički zavod za statistiku. https://data.stat.gov.rs/Home/Result/3104020201?languageCode=sr-Cyrl 4 Među onima koji koriste veštačku inteligenciju (VI): 67,8% proverava sopstveni sadržaj uz pomoć VI. 30% nikada ne proveri sadržaj napravljen od strane VI pre nego što ga prosledi. 30% nikada ne prilagodi sadržaj napravljen od strane VI pre nego što ga prosledi.

Veliki procenat ispitanika je izjavio da ne prepoznaje štetni sadržaj u medijima.

27,6% ispitanika retko ili nikada ne prepoznaje lažne vesti. 51,5% ispitanika retko ili nikada ne prepoznaje teorije zavere. 25% ispitanika retko ili nikada ne prepoznaje govor mržnje. 23,2% ispitanika retko ili nikada ne prepoznaje zagovaranje nečijih ličnih interesa. 71,9% ispitanika nikada ne prepoznaje sadržaj napravljen uz pomoć VI.

Povezano sa tim nalazima, zabeležen je visok uticaj medija na ispitanike.

59,5% retko ili nikada sazna nešto novo u medijima što im promeni stavove o političkom ili društvenom pitanju.

68,6% se svakodnevno ili često oseti preplavljeno informacijama i sadržajima iz medija. 58,3% su svakodnevno ili često izloženi informacijama iz medija koje izazivaju snažne negativne emocije. Kada je reč o proveravanju informacija, iako je 38,4% ispitanika izjavilo da često, a 33,6% povremeno poveruje u informacije iz medija za koje se kasnije ispostavi da nisu tačne, skoro polovina (47,9%) ispitanika retko ili nikada ne proveri informacije iz više izvora, dok 65,8% retko ili nikada ne pročita komentare ispod objave, a 90,3% retko ili nikada ne proveri informacije na nekom sajtu za proveru činjenica. Istinitost informacija iz medija ispitanici u najvećem procentu proveravaju sa nekim koga poznaju, što je izjavilo njih čak 89,7%. Najpozudaniji izvori informacija su za 68,5% ispitanika prijatelji, rodbina i kolege, a mnogo manji procenat ima poverenje u medije. Da je televizija uglavnom ili potpuno pouzdan izvor informacija reklo je 36,9%, slede mesindžeri (32,4%), radio (22%), VI (20,5%), društvene mreže (17,6%), onlajn informativni portali (11,7% ), a najmanje poverenja ispitanici imaju u štampu (3,7%). Od digitalnih veština koje su merene, nešto više od polovine ispitanika ume da isključi praćenje svoje lokacije (55,7%) ili digitalnog ponašanja (52,2%), odnosno da blokira ili izbriše sa liste prijatelja neku osobu (54,9%). Najmanje su razvijene veštine zaštite od štetnog sadržaja koje ne poseduje 86,2%, odmah za tim je nedostatak veština zaštite od onlajn prevare (86%), prijavljivanje govora mržnje (83,9%) i isključivanje memorije veštačke inteligencije (82,5%). Podaci pokazuju da najveći broj ispitanika internet i medije koristi pasivno – primarno za informisanje o važnim temama (94,9%) i povezivanje 65,5%. Ostale aktivnosti su prisutne u mnogo manjem procentu. 5  O istraživanju

Kao deo zahteva Strategije razvoja sistema javnog informisanja u Republici Srbiji za period 2020–2025 (u daljem tekstu: Medijske strategije) istraživanje medijske pismenosti opšte populacije u Srbiji deo je projekta Indeks medijske pismenosti. Ovaj projekat se sprovodi u saradnji i uz podršku Misije OEBS u Srbiji a u cilju realizacije Akcionog plana Medijske strategije, u kojoj je (Odeljak 5.1) istaknut nedostatak kontinuiranog sistemskog i sistematskog merenja medijske pismenosti građana, odnosno kumulativnih efekata različitih indikatora značajnih za njen razvoj. Budući da razvijanje instrumenta za istraživanje medijske pismenosti opšte populacije predstavlja prve korake u ostvarenju Akcionog plana Medijske strategije 2020–20253, najpre je urađena prva radna verzija indeksa medijske pismenosti kako bi se identifikovali najvažniji indikatori i dostupni podaci za njih. Tamo gde nije bilo moguće doći do podataka koji se odnose na ključne dimenzije medijske pismenosti, uključena su kao baterija pitanja u okviru upitnika. Osnovne dimenzije medijske pismenosti koje su pokrivene ovim upitnikom odnose se, prema EAVI modelu4, na individualne kompetencije (pristup i upotreba, odnosno kritička dimenzija) i socijalne kompetencije (komunikacija i participacija). Akcenat je stavljen na kritičku dimenziju medijske pismenosti.

Oslanjajući se na EAVI model medijske pismenosti i njegove osnovne dimenzije koje se odnose na korisnike, formulisana su i istraživačka pitanja:

RQ 1. Na koji način građani Srbije koriste medije (tradicionalne i digitalne) i veštačku inteligenciju?

RQ 2. Kakve kritičke sposobnosti razvijaju kada je reč o medijima i veštačkoj inteligenciji?

RQ.3. Kakve oblike komuniciranja i participacije možemo identifikovati u digitalnom prostoru i koje digitalne veštine građani razvijaju?

Uzorak

U istraživanju medijske pismenosti opšte populacije Republike Srbije uzorak je zasnovan na teoriji verovatnoće. Slučajni metod izbora jedna je od osnovnih pretpostavki donošenja nepristrasnih ocena parametara opšte populacije, jer se očekuje da slučajno odabrane jedinice odražavaju opšte karakteristike opšte populacije. Uzorak je definisan kao dvoetapni stratifikovani reprezentativni uzorak stanovništva Srbije (pošto se smanjivanje grešaka traženih parametara postiže smanjivanjem varijabiliteta između

3 Medijska strategija 2020–2025, Aktivnost 5.1.1. navodi da je potrebno "uspostaviti indeks medijske pismenosti u Republici Srbiji koji je u skladu sa evropskom metodologijom". Izvor: https://www.media.srbija.gov.rs/medsrp/dokumenti/medijska_strategija210_cyr.pdf. 4 https://ec.europa.eu/assets/eac/culture/library/studies/literacy-criteria-report_en.pdf. 6 jedinica na kojima se izvodi istraživanje) koji obuhva 800 ispitanika. Postupak podele naziva se stratifikacija, a dobijeni segmenti nazivaju se stratumi. Osnovne stratume su činile velike teritorijalne celine: Beograd, Vojvodina, i centralna Srbija. U okviru svakog od tih stratuma definisani su urbani i ruralni substratumi što daje ukupno 6 stratuma. Model uzorka omogućava ekspanziju na osnovni skup pa se podaci odnose na stanovništvo Srbije (teritorija Kosova i Metohije nije bila uključena u istraživanje), od 18 do 75 godina, s mogućnošću iskazivanja i prema sociodemografskim obeležjima - prema polu (muškarci i žene), godinama starosti (od 18 do 29 godina, od 30 do 44 godine, od 45 do 59 i od 60 do 75 godina), obrazovanju (nezavršena i završena osnovna škola, srednja škola i viša, visoka škola ili fakultet), i regionu (Beograd, Vojvodina i centralna Srbija). Ovakav uzorak omogućuje zaključivanje na nivou Srbije sa tolerišućom greškom u oceni proporcija od 3,5 procentnih poena.

Upitnik

Semi-standardizovani upitnik za istraživanje medijske pismenosti opšte populacije sadrži pitanja zatvorenog tipa sa izuzetkom jednog pitanja otvorenog tipa i identifikaciona obeležja ispitanika (pol, godine starosti i školska sprema). Kod otvorenog pitanja od ispitanika se očekivalo da sam ponudi odgovor na postavljeno pitanje. U upitniku postoji samo jedno ovakvo pitanje i ono se odnosilo na godinu rođenja ispitanika. Zatvorena pitanja sadrže već ponuđene odgovore.

Za kreiranje upitnika, radi uporedivosti, konsultovana su istraživanja medijske pismenosti opšte populacije u evropskim zemljama (Ofcom, 2024; Oxford, 2019) i zemljama regiona (Paško, 2023; Raičević & Vranjanac, 2023), kao i Republički zavod za statistiku koji je do prošle godine u svojim godišnjim istraživanjima o upotrebi IKT u Srbiji sadržao i pitanja koja su se odnosila na medijsku pismenost. Pored e-veština, ona su obuhvatala i četiri pitanja koja su se odnosila na iskustvo ispitanika sa dezinformacijama – da li su se sa njima sreli, da li su ih proverili, kako su ih proverili i, ako nisu, zašto nisu, kao i set pitanja koji se odnosio na privatnost (Kovačević i dr., 2023). U izveštaju za 2024. godinu grupe pitanja koje su se odnosile na digitalne veštine i zaštitu privatnosti zamenjene su setovima pitanja koji se odnose na industriju - korišćenje pametnih uređaja i zeleni IKT (Josipović i dr. 2024). To znači da u ovom momentu u Srbiji više ne postoji nijedna institucija koja na reprezentativnom uzorku i na godišnjem nivou meri ni minimum ovih važnih kompetencija građana.

Većina studija istraživanja medijske pismenosti započinje setom pitanja o pristupu i načinu na koji građani koriste različite medije. Ta pitanja se odnose na učestalost i dinamiku korišćenja medija, uređaja i interneta, zatim na različite vrste medija i medijskih sadržaja, sa fokusom na informativne medije i društvene mreže. Drugi element medijske pismenosti koji se ponavlja u raznim istraživanjima se odnosi

7 na kritičko mišljenje. Uglavnom sadrži bateriju pitanja koja se odnose na poverenje u medije i često se fokusira na oblast informisanja i dezinformacije. Participacija je treći element medijske pismenosti koji obuhvata i komuniciranje i kreiranje, ali i građanski angažman u analiziranim studijama. Upitnik Internet instituta u Oxfordu (OXI, 2019) ima razrađene baterije pitanja o proizvodnji sadržaja i mnoštvu drugih aktivnosti na internetu. Republički zavod za statitstiku je akcenat stavio na ekonomski aspekt, odnosno na veštine bezbedne kupovine i komunikaciju sa javnom upravom. Hrvatsko istraživanje je takođe akcenat stavilo na veštine proizvodnje sadržaja (Paško, 2023), dok crnogorska studija ovaj deo povezuje sa ključnom temom istraživanja – dezinformacijama (Raičević & Vranjanac, 2023).

U ovom upitniku, težište je na onim dimenzijama medijske pismenosti koje se odnose na kritičko promišljanje i evaluaciju medija i sadržaja, pa su i sva druga pitanja (npr. korišćenje medija) adaptirana u skladu sa tim. Istraživan je način korišćenja različitih medija, kako tradicionalnih tako i digitalnih, uključujući i veštačku inteligenciju koja postaje sve važnija u generisanju, kreiranju i uređivanju sadržaja, kao i proveri informacija. Takođe, istaknute su one aktivnosti koje obuhvataju digitalne veštine, proveru, kreiranje i prosleđivanje sadržaja, kao i ulogu veštačke inteligencije. Iako medijska pismenost obuhvata i popularne formate, zbog ograničenosti prostora i ekonomičnosti koje zahteva CATI metod, oni su pokriveni u manjoj meri. Upitnikom su dominantno obuhvaćeni informativni mediji, veštačka inteligencija i izazovi koje oni otvaraju kada je reč o razvoju medijske pismenosti.

Istraživanje je obavljeno od 15. maja do 15. juna 2025. godine na teritoriji Srbije. Korišćena je tehnika telefonske ankete uz pomoć telefonsko-kompjuterske platforme CATI (Computer Assisted Telephone Interview) koja uz specijalizovane softvere obezbeđuje direktnu kompjutersku podršku različitim fazama istraživanja. Podaci prikupljeni upitnikom su obrađeni u programu za statističku analizu SPSS.

Podaci dobijeni istraživanjem predstavljaju odraz subjektivnih stavova i mišljenja ispitanika. Kada je reč o kritičkom promišljanju, ovakav vid analize je potrebno dopuniti kvalitativnim metodama. Svakako, obuhvatanje svih važnih elemenata i pitanja jednim upitnikom ne samo što je kompleksan zadatak koji zahteva ozbiljne resurse, nego nije uvek ni moguć. Istraživači koji su sudelovali u evaluaciji evropskog modela analize indikatora medijske pismenosti (EC, 2011) to ne preporučuju. „S obzirom na opseg konteksta i ponašanja u vezi s medijskom pismenošću, jednostavna 20-minutna anketa, bez obzira na kvalitet njenog dizajna, ne može pružiti sveobuhvatne podatke potrebne za oblikovanje javnih politika (...) Pokušaj sveobuhvatnog merenja unutar jednog istraživanja nije ni izvodljiv ni preporučljiv.“ Bulger (2012, p. 95). Oni predlažu takozvani modularni pristup, koji se sastoji od dva dela. Jedno istraživanje bi se usredsredilo na osnovna pitanja medijske pismenosti koja se redovno kvantifikuju, dok bi se drugo

8 odnosilo na razvoj rotirajućeg segmenta istraživanja sa fokusom na „specifične komponente medijske pismenosti i mogao bi biti fleksibilan da se prilagodi novim nalazima ili prioritetima“ (ibid.).

Reference

Bilić, Paško. 2023. Medijska pismenost građana Republike Hrvatske. istraživački izvještaj. Institut za razvoj i međunarodne odnose (IRMO). https://gong.hr/wp-content/uploads/2023/07/MOV-Medijska- pismenost-gradana-Republike-Hrvatske.pdf. Bulger, Monica. 2012. Measuring media literacy in a national context: challenges of definition, method and implementation. Media Studies, 3(6), 83-104. http://hrcak.srce.hr/96376. European Commission. 2011. Testing and Refining Criteria to Assess Media Literacy Levels in Europe Final Report. Commissioned by the European Commission Directorate-General for Information Society and Media; Media Literacy Unit. Danish Technological Institute & European Association for Viewers’ Interests. https://eavi.eu/wp- content/uploads/2017/08/study_testing_and_refining_ml_levels_in_europe.pdf- Josipović, Branko, Šutić, Vladimir, Rajčević, Uroš, Minaeva, Ivana. 2024. Upotreba informaciono- komunikacionih tehnologija u Republici Srbiji. Beograd: Republički zavod za statistiku. https://publikacije.stat.gov.rs/G2024/Pdf/G202416019.pdf. Kovačević, Miladin, Šutić, Vladimir, Rajčević, Uroš i Minaeva, Ivana. 2023. Upotreba informaciono- komunikacionih tehnologija u Republici Srbiji. Beograd: Republički zavod za statistiku. https://publikacije.stat.gov.rs/G2023/Pdf/G202316018.pdf. Ofcom. 2024.Adults’ Media Use and Attitudes Report. https://www.ofcom.org.uk/media-use-and- attitudes/media-habits-adults/adults-media-use-and-attitudes-2024-interactive-report. Oxford Internet Institute – OXI. 2019. Oxford Internet Survey 2019 General Questionnaire. https://oxis.oii.ox.ac.uk/wp-content/uploads/sites/16/2019/11/OxIS-Questionnaire-2019.pdf. Raičević, Vladimir & Vranjanac, Marija. 2023. Medijska pismenost u Crnoj Gori. Istraživanje javnog mnenja. Ipsos. https://www.osce.org/files/f/documents/8/2/564892.pdf. Savet za štampu. 2020. Leksikon medijske pismenost. https://savetzastampu.rs/wp-content/uploads/2020/12/3-LEKSIKON-MEDIJSKE-PISMENOSTI.pdf.

9 SOCIO-DEMOGRAFSKE KARAKTERISTIKE ISPITANIKA

Malo više od polovine ispitanih su žene, što odražava polnu strukturu stanovništva Republike Srbije. U uzorku je 409 žena (51,1%) i 391 pripadnik muškog pola (48,9%). Među ispitanicima je bilo 31,9% onih koji imaju 60+ godina. Skoro četvrtina ispitanika (24,3%) ima između 45 i 59 godina, dok između 30 i 44 godina ima 27,6 učesnika ankete. Najmanji procenat je onh koji imaju 18-29 godina (16,2%).

Tabela 1: Pol i ispitanika (n i %) Tabela 2: Uzrast ispitanika:

Pol ispitanika Broj % Uzrasna grupa Broj % Muški 391 48,9 18 - 29 130 16,2 Ženski 409 51,1 30 - 44 221 27,6 Ukupno: 800 100 45 - 59 194 24,3 60 - 75 255 31,9 Ukupno: 800 100

Kada je o obrazovanju reč, više od polovine ispitanika ima završenu srednju školu (55,9%), petina je visokoobrazovanih (20,5%), 6,5 odsto ih je sa nezavršenim fakultetom ili višom školom, i 17,1 odsto sa završenom ili nezavršenom osnovnom školom (Tabela 2).

Tabela 3: Obrazovanje ispitanika (n i %)

Poslednja završena školska sprema Broj % Bez i OŠ 137 17,1 SSS 448 55,9 Nezavršen fakultet/viša/visoka škola 52 6,5 Viša/fakultet 163 20,5 Ukupno: 800 100

Takođe, najveći broj ispitanika je zaposlen (50,9%) ili u penziji (17,9%). Nezaposlenih je 5,9%, slično kao onih koji studiraju (6,3%). Među ispitanicima je i 9% osoba koje se brinu o domaćinstvu, odnosno 10% onih koji se bave poljoprivredom (Tabela 3).

Tabela 4: Radni status ispitanika (n i %)

Radni status Broj % Zaposlen/a 408 50,9 Nezaposlen/a 47 5,9 Penzioner/ka 143 17,9 Student/kinja 50 6,3 Poljoprivrednik/ca 80 10 Osoba koja brine o domaćinstvu 72 9 Ukupno: 800 100

10 Najveći broj ispitanika – skoro polovina - živi u Centralnoj Srbiji (48,3%), a slede ispitanici iz Vojvodine (26,3%) i Beograda (25,4%). Više od polovine (58,9%) živi u gradskim sredinama, dok je onih koji žive u ruralnim naseljima 41,1%.

Tabela 5: Regioni u kojima žive ispitanici (n i %): Tabela 6: Tip naselja u kojima ispitanici žive (n i %):

Region Broj % Centralna Srbija 386 48,3 Tip naselja Broj % Beograd 204 25,4 Gradsko (urbano) 471 58,9 Vojvodina 210 26,3 Vangradsko (ruralno) 329 41,1 Ukupno: 800 100 Ukupno: 800 100

Većina ispitanika (39,8%) živi u domaćinstvima čija mesečna primanja se kreću od 60.000 do 150.000 dinara. Značajan broj – više od petine, tj. 22,4% - živi u domaćinstvu koje mesečno prihoduje do 60.000 dinara. Od 150.000 do 200.000 dinara mesečno5 su primanja o kojima se izjasnilo 18,3% ispitanih, a samo 8% je onih čija primanja na nivou domaćinstva iznose preko 200.000 dinara. Gotovo desetina (11,5%) anketiranih nije želela da se izjasni o mesečnim prihodima njihovih domaćinstva.

Tabela 7: Prosečna primanja na nivou domaćinstva

Prosečna mesečna primanja na nivou Broj % domaćinstva Do 60.000 dinara 179 22,4 Od 60.000 do 150.000 dinara 319 39,8 Od 150.000 do 200.000 dinara 146 18,3 Preko 200.000 dinara 64 8 Ne želim da se izjasnim 92 11,5 Ukupno: 800 100

5 Prosečna neto zarada u Srbiji u maju 2025. godine iznosila je 107.705 dinara, dok je medijalna neto zarada bila 84.408 dinara: https://stat.gov.rs/en-us/vesti/statisticalrelease/?p=17056 11  Rezultati istraživanja

I PRISTUP I KORIŠĆENJE MEDIJA I VEŠTAČKE INTELIGENCIJE

U ovom delu istraživanja merena je učestalost upotrebe određenog medija i platformi veštačke inteligencije. Prva grupa pitanja u upitniku odnosila se na učestalost upotrebe određenog medija na osnovu sopstvene procene ispitanika. Na pitanje koliko često koriste određeni medij ispitanici su odgovarali birajući jedan od ponuđenih odgovora na petočlanoj skali, prema kojoj je 1-nikada, 2-retko, 3-povremeno, 4-često i 5-svakodnevno (Videti Grafikon 1).

Grafikon 1: Učestalost korišćenja medija (tradicionalnih i digitalnih) i veštačke inteligencije (VI), (u %)

90 80 70 60 50 40 30 20 10 0 STRIMING Platforme ONLAJN PLATFORME MUZIČKE veštačke DRUŠTVENE TELEVIZIJU RADIO ŠTAMPU INFORMATIV za gledanje STRIMING MESINDŽERE inteligencije MREŽE NE PORTALE serija i PLATFORME (npr. filmova ChatGPT) 1-nikada 3.3 33.7 77.4 43.5 64.3 73.3 42.5 18.5 74.6 2-retko 14 25.1 14 16.9 12.8 5.8 9.1 5.3 5.9 3-povremeno 26.5 27.4 6 24.6 17.1 15 19.2 20 12.9 4-često 20.9 10.1 1.3 11.9 4.8 5 16.8 26.4 5.3 5-svakodnevno 35.3 3.6 1.3 3.1 1 0.9 12.4 29.8 1.3

Prema podacima dobijenim ovim istraživanjem, dominantni medij kada je reč o učestalosti upotrebe je televizija – više od polovine ispitanika je gleda ili često (20,9%) ili svakodnevno (35,3%). Takođe, samo je 3,3% ispitanika reklo da nikada ne gleda televiziju. Gotovo jednako su u upotrebi i aplikacije (mesindžeri) za komuniciranje i razmenu poruka (Viber, WhatsApp i sl.) – 29,8% ispitanih ih koristi svakodnevno, 26,4% često, oko četvrtine ispitanih ih retko (5,3%) ili povremeno koristi (20%), a oko petine (18,5%) nikada ne koristi ove aplikacije.

Evidentno je, takođe, da su tradicionalni mediji poput štampe i radija značajno ustupili prostor televiziji. Radio nikada ne sluša trećina (33,7%) ispitanika, dok svaki četvrti (25,1%) to čini retko, a tek 3,6% svakodnevno. Ipak, medij koji je najmanje zastupljen među korisnicima je štampa, koju nikada ne čita čak 77,4% ispitanika. Ovome treba dodati i 14% onih koji to čine retko, što ostavlja tek nešto iznad 2,5% onih koji to čine često ili svakodnevno. 12 Noviji oblici informisanja putem informativnih onlajn portala zastupljeni su kod nešto više od polovine ispitanih, ali ih 43,5% uopšte ne koristi. Njih 16% to čini retko, a četvrtina ispitanika (24,6%) povremeno. Preko portala se često informiše 11,9% ispitanika, a 3,1% svakodnevno, što je gotovo jednako procentima lojalnijih korisnika radija.

Striming platforme ne koristi veliki broj ispitanih građana – 73,3% ispitanika uopšte ne koristi muzičke striming platforme (npr. Spotifaj, Dizer), dok 64,3% ne koristi ovu vrstu servisa za gledanje serija i filmova (npr. Netflix, Max, Amazon Prajm, Dizni +). Samo 1% to čini svakodnevno, oko 5% često, a najveći broj onih koji koriste ovu vrstu sadržaja to čini povremeno – 17,1% ispitanika gleda, odnosno 15% sluša sadržaje preko posebnih striming servisa.

Društene mreže svakodnevno ili često koristi 12,4% odnosno 16,8% ispitanika. Svaki peti (19,2%) ispitanik ih koristi povremeno, retko 9,1% dok ih ne koristi 42,5 odsto. Važno je istaći da ovaj podatak ne govori o broju ispitanika koji imaju profile na društvenim mrežama nego isključivo o njihovoj učestalosti upotrebe u poslednjih mesec dana.

Zanimljivo je da je korišćenje veštačke inteligencije zastupljeno u sličnoj meri kao korišćenje striming platformi. Većina ispitanika je nikada nije koristilo (74,6%), 5,9% to čini retko, 12,9% povremeno, a tek 5,3%, odnosno 1,3% ispitanika je koristi često ili svakodnevno. S obzirom na to da su mogućnosti strimovanja sadržaja, ali i aplikacije veštačke inteligencije dostupne tek odskora, biće zanimljivo videti koliko brzo će njihova upotreba biti popularizovana.

KORIŠĆENJE VEŠTAČKE INTELIGENCIJE

Ispitanicima je postavljena i posebna grupa pitanja (Grafikon 2) sa ciljem da se izmeri diverzitet upotrebe veštačke inteligencije. Tri četvrtine ispitanika (oko 75%) je odgovorilo da ne koristi veštačku inteligenciju, dok je 25,4% ispitanika reklo da je koriste.

Oni koji je koriste, to rade u većoj meri za proveru i poboljšanje sadržaja koji su prethodno napravili (67,8%) kao i za istraživanje i dolaženje do ideja pre nego što kreiraju sadržaj (56,4%). Manje od polovine ispitanika koji koriste VI, to rade za traženje uputstava za kreiranje i koncipitanje sadržaja (46,2%), odnosno za kreiranje sadržaja (45,7%), dok je najmanji procenat njih koristi sa ciljem pravljenja strategije za plasman i promociju sadržaja (38,4%).

13 Grafikon 2: Razlozi za korišćenje veštačke inteligijencije među ispitanicima koji su rekli da je koriste (u %) 80 67.8 70 61.6 54.3 56.4 53.8 60 45.7 43.6 46.2 50 38.4 40 32.2 30 20 10 0 kreiranje sadržaja (tekst, istraživanje i dolaženje do uputstva za kreiranje i pravljenje strategije za proveru i poboljšanje fotografija, video)? ideja pre nego što koncipiranje sadržaja? plasman i promociju sadržaja koji ste kreirate sadržaj? sadržaja? prethodno napravili?

da ne

Takođe, ispitanicima su postavljena i pitanja koliko često sadržaje koje je kreirala veštačka inteligencija prilagode, odnosno provere pre nego što ih proslede dalje (Grafikon 3). Oko trećine ispitanika je kazalo da sadržaj prilagode uvek (32%) ili nikada (30,1%), odnosno provere ga uvek (28,2%) ili nikada (30) pre nego pre nego što ga proslede dalje. Obe aktivnosti čini često otprilike svaki peti ispitanik, dok svaki deseti to čini retko.

Grafikon 3: Učestalost provere u olrilagođavanja sadržaja kreiranog VI (u %) . Kada koristite VI koliko često uradite sledeće:

40 30.1 32 30 28.2 30 18.7 20 20 12.7 10.1 9.1 9.1 10

0 PRILAGODITE pre nego što ga prosledite dalje? PROVERITE pre nego što ga prosledite dalje?

nikada retko povremeno često uvek

II KRITIČKO ČITANJE

Ovaj deo pitanja bio je usmeren na procenu sagovornika o sopstevnim sposobnostima za prepoznavanje različitih sadržaja u medijima bez obzira da li je reč o nekom obliku dezinformisanja ili manipulacijama koje su proizvedene korišćenjem veštačke inteligencije. Ispitanici su imali opciju da na pitanja odgovore na petostepenoj skali: 1-nikada, 2-retko, 3-povremeno, 4-često i 5-uvek (Grafikon 4).

Ispitanici najveću sigurnost u sopstvenu procenu imaju kada je reč o prepoznavanju govora mržnje i zagovaranja ličnih interesa u medijima. Tako 7% anketiranih kaže da ih uvek sa sigurnošću prepoznaju, a preko 40% da je to često slučaj – u nešto većem broju kada je reč o zagovaranju ličnih interesa (43,9%) naspram govora mržnje (40,4%). Sa manjom sigurnošću ispitanici govore o lažnim vestima i teorijama zavere - u veoma malom procentu mogu sa sigurnošću da potvrde svoje sposobnosti procene ovakvih 14 sadržaja u svakoj situaciji (1,4% uvek prepoznaje lažne vesti, a 0,6% teorije zavere). Međutim, kada je o lažnim vestima reč, više od trećine ispitanika ih sa sigurnošću prepoznaje često (33%) ili povremeno (38%). S druge strane, u mnogo manjem broju su sigurni da u medijima prepoznaju teorije zavere – 13% ih je često sigurno da ih prepoznaju, a 34,9% povremeno sa sigurnošću prepoznaju ove sadržaje. Ispitanici su najmanje sigurni u svoje sposobnosti da dovoljno često prepoznaju sadržaj generisan pomoću veštačke inteligencije – čak 71,9% je reklo da nikada nije sigurno da ih prepoznaje.

Grafikon 4: Prepoznavanje sadržaja (u %)- Koliko često ste sigurni da ste u medijima prepoznali :

80 71.9

60 34.9 40.4 43.9 38 40 33 31.9 27.6 25.8 17.8 19.6 20 9.8 13 10.614.4 7 9.6 13.6 7.1 10.413.6 1.4 0.6 3.7 0.4 0 lažne i namerno teorije zavere? govor mržnje? zagovaranje nečijih ličnih sadržaj napravljen uz obmanjujuće vesti? interesa? pomoć veštačke inteligencije?

1-nikada 2-retko 3-povremeno 4-često 5-uvek

UTICAJ INFORMISANJA

Upitnikom je adresiran i uticaj koji mediji imaju na osećanja i stavove korisnika (Grafikon 5). Upitani o tome koliko često su u medijima saznali nešto novo što im je promenilo stavove o političkom ili društvenom pitanju, više od polovine ispitanika je odgovorilo da se to dešavalo retko (32,1%) ili nikada (27,4%). Njih 27,9% je reklo da se to povremeno dešavalo, dok 11,5% kaže da mediji često imaju ovakav uticaj na njih.

Mediji imaju snažan uticaj kada je reč o emotivnom odgovoru koji izazivaju. Većina ispitanika odgovorila je da se često (41,7%) ili svakodnevno (26,9%) oseća preplavljeno emocijama i sadržajima iz medija. Samo 2,6% kaže da to nikada nije doživelo, a 8,5% da se to retko dešavalo. Preplavljenost emocijama zbog medijskih sadržaja povremeno doživljava petina ispitanika (19,9%).

Takođe, više od polovine ispitanika je reklo da medijski sadržaji često (47,9%) ili svakodnevno (10,4%) kod njih izazivaju snažne negativne emocije, poput tuge, besa, stida, gađenja, straha, itd. Povremeno se tako oseća značajan broj ispitanika – njih 31,8%, a samo 2,4% kaže da se nikada nije tako osećalo zbog nečega što je videlo ili čulo u medijima.

15 Grafikon 5: Uticaj medija na korisnike (u %) – Koliko često ste u medijima:

60

47.9 50 41.7 40 32.1 31.8 27.4 27.9 26.9 30 19.9 20 11.5 10.4 8.5 7 10 0.8 2.6 2.4 0.3 0.4 0.5 0 saznali nešto NOVO što vam je osetili se PREPLAVLJENO bili izloženi informacijama iz medija koje PROMENILO stavove o političkom ili INFORMACIJAMA i sadržajima iz medija? su Vam izazvale SNAŽNE NEGATIVNE društvenom pitanju? EMOCIJE?

1-nikada 2-retko 3-povremeno 4-često 5-svakodnevno 77-ne znam

PROVERA SADRŽAJA Pitanja u vezi sa praksom proveravanja informacija, odnosno medijskih sadržaja, takođe su skalirana od 1 do 5 (1 - nikada, 2-retko, 3-povremeno, 4-često, 5-uvek). Ispitanici su odgovarali na pojedinačna pitanja o mogućim načinima proveravanja sumnjivih informacija na koje nailaze u medijima – od provere iz više različitih medijskih izvora, preko čitanja komentara, konsultovanja sa osobama koje poznaju, do korišćenja fakt-čeking resursa ili veštačke inteligencije (Grafikon 6).

Ispitanici najređe informacije proveravaju pomoću sajtova ili aplikacija za proveru činjenica – 84% kaže da to nikada ne radi, njih 6,3% retko, 5,8% povremeno, a samo 3,1% često, odnosno 0,8% svakodnevno koristi ove izvore. Takođe, mali broj ispitanika koristi veštačku inteligenciju za ovu svrhu. Svakodnevno je koristi tek 0,3%, često 6,1%, povremeno 11%, 6,8% retko, a 75,5% nikada nije koristilo veštačku inteligenciju za proveru sumnjivih informacija iz medija.

Ako im je neka informacija iz medija sumnjiva, ispitanici je najčešće proveravaju konsultovanjem sa osobama koje poznaju ili drugih medijskih izvora, s tim što u ovom slučaju poznanici i bliske osobe imaju blagu prednost u odnosu na medije. Kada je reč o proveri informacija sa osobama koje poznaju, samo 10,3% je reklo da to nikada nije radilo, a 32,6% da to retko čini. Povremeno tako proverava činjenice 35,5% ispitanika, 18,8% to čini često, a 2,8% svakog dana. Poređenja radi, druge medijske izvore konsultuje svaki peti ispitanik i to pre često (18,5%) svakodnevno (1,8%). Povremeno to radi 31,5% njih, retko 26%. Više od petine ispitanika (21,9%) nikada ne proverava sumnjive informacije iz više izvora.

16 Komentare na sumnjiv sadržaj ili informaciju nikada ne čita polovina ispitanika, dok nih samo 1,4% to čini svakodnevno. Među onima koji to čine, jedna petina (20,5%) to čini povremeno, 15,8% retko, i 12,3% je izjavilo da to radi često.

Pitanjima u vezi s proverom sumnjivih sadržaja i informacija iz medija dodato je pitanje o ličnom iskustvu sa lažnim ili obmanjujućim vestima. Na pitanje koliko često se dešavalo da poveruju u neku informaciju u medijima za koju se kasnije ispostavi da nije tačna, 38,4% ispitanika je reklo da im se to često dešava, a 33,6% da se to dešava povremeno. Petini ispitanih (20,8%) se retko dešava da poveruju u lažne vesti, a 6,4% anketiranih tvrdi da im se to nikada nije desilo.

Grafikon 6: Lično iskustvo ispitanika sa lažnim ili obmanjujućim vestima (u %) – Koliko često sumnjiv sadržaj ili informacije iz medija:

proverite na poverujete u nekom sajtu ili proverite pitajući proverite sa informacije iz proverite iz više pročitate aplikaciji za veštačku nekim koga medija, za koje se izvora? komentare? proveru činjenica inteligenciju? poznajete? kasnije ispostavi (fakt-čeking)? da nisu tačne? 1-nikada 21.9 50 84 75.5 10.3 6.4 2-retko 26 15.8 6.3 6.8 32.6 20.8 3-povremeno 31.5 20.5 5.8 11 35.5 33.6 4-često 18.5 12.3 3.1 6.1 18.8 38.4 5-svakodnevno 1.8 1.4 0.8 0.3 2.8 0.4 77-ne znam 0.3 0 0 0.3 0 0.4

1-nikada 2-retko 3-povremeno 4-često 5-svakodnevno 77-ne znam

KREDIBILITET IZVORA Kada je reč o pouzdanosti izvora informisanja, cilj je bio da se izmeri kome ispitanici poklanjaju najveće poverenje - nekom od medija, veštačkoj inteligenciji ili porodici, prijateljima i kolegama. Ispitanici su birali odgovore na skali od 1 do 5, gde bi jedinicom ocenili izvor koji uopšte nije pouzdan, a peticom onaj koji je u potpunosti pouzdan. Ispitanicima je bilo omogućeno i da odgovore da nisu u mogućnosti da procene pouzdanost izvora (Grafikon 7).

17 Najviše poverenja ispitanici imaju u osobe koje poznaju – svoje prijatelje, kolege, porodicu – preko 68% ispitanika smatraju ih u potpunosti (33,1%) ili uglavnom pouzdanim (35,4%) izvorom informacija, a 22,9% misli da i jesu i nisu pouzdani.

Grafikon 7: Pouzdanost izvora informacija (u %)

1- uopšte nije pouzdana 2 - uglavnom nije pouzdana 3 - i jeste i nije pouzdana 4 - uglavnom je pouzdana 5 - u potpunosti je pouzdana 77 - ne znam, ne mogu da procenim

77.9

74.6 59.6

50.4

43.5

35.4 34.4

33.1 32.3 27.5

26.6 23.8 23.3

22.9 22.5 20.3

14.9 14.5 14.5

14.1 14.1 13.6

10.9 10.1 9.4 7.5

7.3 6.7

6.1 5.8

5.6 4.9

4.4

4.1 3.5

3.5 3.3

2.6

2.4 2.3 1.6 1.1 0.6

0.6

0.4 0.4

0.1 1

Televizija Radio Štampa Informativni Društvene Mesindžeri Veštačka Porodica, portali mreže inteligencija prijatelji, kolege

S obzirom na to da je veliki broj ispitanika odgovorio da ne može da proceni pouzdanost pojedinih vrsta medija, moguće je da ovakvi odgovori korespondiraju sa učestalošću njihovog korišćenja – ispitanici se osećaju kompetentnije da procene pouzdanost medija koji češće koriste i bolje razumeju i obrnuto. Na primer, 74,6% ispitanika ne može da proceni koliko je veštačka inteligencija pouzdana, a – zanimljivo - slična je situacija i sa štampom, gde čak 77,6% ispitanih ne može da kaže koliko je ovaj medij kredibilan izvor informacija. Pitanje o kredibilitetu radija takođe beleži čak 59,6% onih koji ne mogu da procene njegovu pouzdanost. Poređenja radi, pouzdanost televizije kao izvora informisanja ne može da proceni samo 3,5% ispitanika. Ispitanici koji su se izjašnjavali o kredibilitetu medija u najvećem broju su ambivalentnog stava i smatraju ih i pouzdanim i nepouzdanim, što svakako može da se dovede u vezu i sa opštošću pitanja i raznolikošću medija koji pripadaju istoj vrsti.

O pouzdanosti televizije izjasnilo se 96,5% ispitanika. U potpunosti ili uglavnom pouzdanim izvorom smatra je 36,9% ispitanika naspram 25,2% onih koji je smatraju nekredibilnim ili uglavnom nekredibilnim izvorom informacija. Pritom, 34,4% je smatra i pouzdanom i nepouzdanom. O radiju se

18 izjasnilo oko 40% anketiranih, 14,5% ga smatra i pouzdanim i nepouzdanim, ali ga mnogo veći broj ispitanika smatra pouzdanim (ukupno 22%) nego nepouzdanim (ukupno 3,9%).

Pri ocenjivanju informativnih portala, najveći broj ispitanika je ambivalentan (27,5%). Pouzdanim izvorom ih smatra 11,7% u odnosu na 10,4% onih koji nemaju poverenja u njih. Više od polovine nije moglo da proceni kredibilitet portala kao izvora informacija. Najmanji broj ispitanika se izjasnio o pouzdanosti štampe – svega 22,1%, ali ako u obzir uzmemo odgovore onih koji se jesu izjasnili, štampa je procenjena kao najmanje kredibilan izvor informacija, pa je kao nepouzdanu ili uglavnom nepouzdanu vidi 11,7% naspram svega 3,7% onih koji je smatraju kredibilnim izvorom.

Vrlo je ujednačen odnos među procentima onih koji veruju (17,6%) i onih koji ne veruju (16,4%) u kredibilitet društvenih mreža kao izvora informacija, ali je veći broj onih koji ih smatraju i pouzdanim i nepouzdanim (22,5%), odnosno ne mogu da procene njihovu kredibilitet (43,5%). Aplikacije za razmenu poruka i veštačka inteligencija ocenjuju se kao kredibilnije od društvenih mreža. Međutim, o pouzdanosti mesindžera se izjasnio mnogo veći broj ispitanika – preko tri četvrtine ispitanika (76,2%) što korespondira sa procentom onih ispitanika koji su izjavili da koriste ove aplikacije svakodnevno, često ili makar povremeno. Od toga, 32,3% ih smatra i pouzdanim i nepouzdanim, 11,5% nepouzdanim, a 32,4% pouzdanim ili uglavnom pouzdanim izvorom informacija. Iako je veštačku inteligenciju procenio manji broj ispitanika (ali više nego štampu!), najmanje ih je bilo ambivalentnih (4,1%) u stavu prema ovom izvoru, a samo 0,8% ga smatra nepouzdanim. Petina ispitanika (20,5%) je ocenila veštačku inteligenciju kao kredibilan izvor informacija.

ULOGA INSTITUCIJA Važan segment medijske i informacione pismenosti je stav o ulozi regulatornog tela za medije, kao i razumevanje razlike između javnih medijskih servisa i komercijalnih medija. Ispitanicima su ponuđene četiri tvrdnje na koje je trebalo da odgovore na skali od 1 do 5, gde jedinicom izražavaju potpuno neslaganje, a peticom potpuno slaganje sa datim iskazom. Prve dve tvrdnje u vezi su s regulatornim telom za medije i odnose se na opseg njegovog rada i svrsishodnost obraćanja građana regulatornom telu u slučaju neprofesionalnog rada medija (Grafikon 8).

19 Grafikon 8: Uloga REM-a (u %)

81.4

48.2

12.9 18.4 8.9 7.8 7.1 12 0.8 0.1 1 1.4

Zadatak REM-a je da prati rad svih medija, opominje i kažnjava Ima smisla da se građani žale REM-u na neprofesionalan rad kada emituju neprimerene sadržaje, nezavisno od političkih i medija ili na konkretan štetni sadržaj koji emituju. ekonomskih uticaja.

1- uopšte se ne slažem 2 - uglavnom se ne slažem 3 - i slažem i ne slažem se 4 - uglavnom se slažem 5 - u potpunosti se slažem 77 - ne znam

Skoro svi ispitanici se slažu (8,9% uglavnom, a 81,4% u potpunosti) sa tvrdnjom da je zadatak REM-a da prati rad svih medija, opominje i kažnjava kada emituju neprimerene sadržaje, nezavisno od političkih i ekonomskih uticaja. Samo po 1% ispitanika se ne slaže sa ovim ili je ambivalentnog stava. Međutim, kada je reč o reagovanju građana povodom neprofesionalnog rada medija, odgovori su donekle drugačiji. Na pitanje da li ima smisla da se građani žale REM-u na neprofesionalan rad medija ili na konkretan štetni sadržaj koji emituju, oko 60% njih se slaže - 12,9% uglavnom i 48,2% potpuno, dok je 12% neodlučno. Samo 8,5% se uglavnom ili uopšte ne slaže sa ovom tvrdnjom.

Druge dve tvrdnje su u vezi s odgovornošću i obavezama medija, kako komercijalnih, tako javnih medijskih servisa (Grafikon 9).

Grafikon 9: Odgovornost komercijalnih mediji i javnog medijskog servisa (u %)

67.8 52

14.3 13.8 12.6 5.4 6.1 8.4 4.4 8.1 5 2.1

Komercijalni mediji ne moraju da budu društveno odgovorni zato što su Javni medijski servisi (RTS, RTV) treba da imaju viši nivo društvene privatni. odgovornosti od privatnih medija.

1- uopšte se ne slažem 2 - uglavnom se ne slažem 3 - i slažem i ne slažem se 4 - uglavnom se slažem 5 - u potpunosti se slažem 77 - ne znam

Da javni medijski servisi treba da imaju viši nivo odgovornosti od privatnih medija misli preko 80% ispitanika (12,6 se uglavnom, a 67,8% potpuno slaže), svega 6,5% ne misli tako, a 8,1% se i slaže i ne slaže. Na pitanje da li su komercijalni mediji lišeni društvene odgovornosti zato što su u privatnom vlasništvu, većina ispitanika odgovara potpunim (52%) ili uglavnom neslaganjem (14,3%), a 13,8% se i

20 slaže i ne slaže. Gotovo svaki deseti ispitanik (11,5% ) se slaže (uglavnom ili potpuno) da komercijalni mediji ne moraju da budu društveno odgovorni zato što su privatni.

III PARTICIPACIJA

Ovaj deo istraživanja je fokusiran na samoprocenu ispitanika o njihovim digitalnim veštinama, ali i na korišćenje medija za različite aktivnosti kao što su interesno povezivanje sa dugim korisnicima i učestvovanje u društvenom životu.

DIGITALNE VEŠTINE Deo istraživanja posvećen digitalnim veštinama daje poseban uvid u razvijenost različitih znanja i iskustava pri upotrebi digitalnih informacionih i komunikacionih tehnologija, poput interneta, mobilnih telefona, računara, korišćenja različitih softverskih aplikacija namenjenih komunikaciji i razmeni poruka, društvenih mreža, itd. sa fokusom na komunikaciji, participaciji, kreiranju sadržaja, ali i zaštiti od nepoželjnih sadržaja i komunikacije. Ispitanicima je postavljen niz pitanja o praksama obezbeđivanja privatnosti i zaštite od štetnog sadržaja koja su zahtevala potvrdan ili odričan odgovor (Grafikon 10).

Prema dobijenim podacima, više od polovine ispitanika zna kako da isključi praćenje svoje lokacije (55,7% je odgovorilo sa „da“ na ovo pitanje, naspram 44,3% koji su rekli da ne umeju) ili da blokira ili izbriše sa liste prijatelja neku osobu (54,9% ume, a 45,1% ne ume). Slično, malo više od polovine ispitanika (52,2%) zna kako da isključi praćenje svog digitalnog ponašanja, npr. istoriju pregledanja.

Nedostatak veština je primetniji u nekim drugim praksama, koje se odnose na zaštitu ličnih podataka i zaštitu od štetnih sadržaja. Na primer, 48,4% korisnika internet kažu da umeju da procene u kojim slučajevima smeju da podele svoje privatne podatke, ali više od polovine kaže da ne ume (51,6%). Takođe, značajno manji je broj – samo 14% - onih koji tvrde da znaju kako da se zaštite od onlajn prevare (krađe identiteta, spema, fišinga, iznude novca i sl.). Samo 17,5% ispitanika ume da isključi memoriju veštačke inteligencije. Kada je reč o zaštiti od štetnog sadržaja, samo 13,8% ispitanika kaže da ume da se zaštiti od sadržaja poput eksplicitnog nasilja, govora mržnje, pornografije, vređanja, ponižavanja, i sličnog. Slično su odgovorili i na pitanje da li umeju da prijave govor mržnje na društvenim mrežama – 16,1% zna kako to da uradi.

21 Grafikon 10: Digitalne veštine (u %)

Da li znate kako da blokirate ili izbrišete sa liste neku osobu/prijatelja? 54.9 45.1 Da li znate kako da se zaštitite od onlajn prevare (krađa ID, spam, fišing, iznuda novca...)? 14 86 Da li znate kako da se zaštitite od sadržaja koji je štetan za Vas (eksplicitno nasilje, govor mržnje, pornografija, vređanje, ponižavanje...)? 13.8 86.2 Da li umete da procenite u kojim slučajevima smete da podelite svoje privatne podatke? 48.4 51.6 Da li umete da isključite memoriju veštačke inteligencije? 17.5 82.5 Da li umete da isključite praćenje svog digitalnog ponašanja (istoriju pregledanja..)? 52.2 47.8 Da li umete da isključite praćenje svoje lokacije? 55.7 44.3 Da li umete da prijavite govor mržnje na društvenim mrežama? 16.1 83.9

0 10 20 30 40 50 60 70 80 90 100

1 – umem 2-ne umem ili ne znam da li umem

AKTIVNOSTI Kada je reč o aktivnostima, podaci ukazuju na primarno korišćenje interneta za informisanje i sporadičnu razmenu mišljenja i povezivanje sa drugim ljudima koji dele iste interese ili brige (Grafikon 11).

Grafikon 11: Najvažnije aktivnosti na internetu (u %)

86.6 81.7 66.3 51.6 29.6 31.1 34.5 28.3 23.9 20.8 20.6 22.7 14.3 13.1 12 16.1 5.1 4.8 9.1 7.3 6.4 5.9 0.5 0.9 0.4 0.1 3.3 1.9 1.1

Informisanje o Povezivanje sa Slobodno iskazivanje Zagovaranje Pozivanje na Učestvovanje ili važnim temama ljudima koji imaju svog mišljenja o određene ideje i odgovornost javne organizovanje (društvenim, slične interese ili važnim pitanjima? ciljeve? ličnosti? događaja, protesta ili političkim, itd.)? brige? kampanje?

1-nikada 2-retko 3-povremeno 4-često 5-svakodnevno 77-ne znam

Više od polovine ispitanika internet koristi da se informiše o važnim društvenim, političkim i drugim pitanjima (31,1% često ili 20,6% svakodnevno, oko 30% povremeno), a samo 5% za redovno iskazivanje sopstvenog mišljenja ili stava. Povremeno se s drugima povezuje 28,3% ispitanika, a više od polovine to čini retko ili nikada (ukupno 58,4%). Skoro tri četvrtine nema običaj da slobodno iskazuje svoje mišljenje u onlajn svetu – a više od polovine (51,6%) to ne čini nikada.

Veoma mali broj ispitanika internet vidi kao platformu za zagovaranje ideja i ciljeva ili organizovanje širih društvenih aktivnosti, protesta, kampanja. Zagovaranjem ciljeva tek povremeno se bavi 14,3% 22 ispitanika (3,3% to čini često), dok retko to čini 16,1%. Njih 66,3% nikada nije koristilo medije i internet u ove svrhe. Kada je reč o učestvovanju ili organizovanju događaja, čak 86,6% nikada nije internet i medije koristilo sa tim ciljem, a 6,4% to čini retko, 5,9% povremeno, a samo 1,1% često. Internet se ne koristi ni za pozivanje javnih ličnosti na odgovornost – 9,1% ispitanih to retko praktikuje, a čak 81,7% nikada. Samo povremeno to radi 7,3% ispitanika, a tek 1,9% kaže da to često čini.

Zaključak i preporuke

Podaci iz istraživanja jasno pokazuju da građani Srbije pretežno gledaju televiziju, a u manjoj meri prate vesti na onlajn portalima, dok ostale tradicionalne medije – radio i štampu – zanemaruju. Posebno je uočljivo nisko poverenje koje su ispitanici izrazili prema štampi i internet portalima. Otprilike četvrtina svih ispitanika koristi veštačku inteligenciju (oko 25%), a najčešće je koriste radi provere sadržaja koji su sami kreirali ili za istraživanje pre kreiranja sadržaja.

Istraživanje pokazuje da dve trećine ispitanika najviše poverenja ima u kolege, prijatelje, poznanike i članove porodice, a manje u medije. Veoma mali broj ispitanika koristi alate za proveru činjenica i u najvećoj meri se oslanja na poznanike pri proveri tačnosti medijskih informacija. Iako su sigurni da prepoznaju sadržaj kao što su govor mržnje ili lična promocija, nisu sigurni kada su u pitanju lažne vesti i teorije zavere—samo vrlo mali procenat može sa sigurnošću reći da uvek može da prepozna takav sadržaj (1,4% za lažne vesti i 0,6% za teorije zavere). Ovo pokazuje nedostatak razvijenih veština kritičkog medijskog čitanja, iako se većina oseća preplavljeno medijskim sadržajem. Najveći problem je i dalje sa veštačkom inteligencijom, gde su ispitanici najmanje sigurni—71,9% je izjavilo da nikada nisu sigurni da li mogu da prepoznaju sadržaj koji je generisala veštačka inteligencija.

Ispitanici internet koriste pasivno, uglavnom radi dobijanja informacija i povezivanja sa drugima putem aplikacija za poruke i društvenih mreža. Samo 5% redovno koristi internet za izražavanje sopstvenih stavova ili mišljenja, više od 60% nikada ga nije koristilo kao platformu za zagovaranje, a više od 86% ga nikada nije iskoristilo da učestvuje ili organizuje neki događaj.

Nedostatak veština je vidljiv i kod zaštite ličnih podataka i zaštite od štetnog sadržaja. Nešto više od polovine ispitanika zna kako da isključi praćenje lokacije (55,7%) ili digitalnih aktivnosti (52,2%) i kako da blokira ili ukloni nekoga sa liste prijatelja (54,9%). Najmanje razvijene veštine su u zaštiti od štetnog sadržaja (nedostaju kod 86,2%), kao i u zaštiti od onlajn prevare (86%), prijavljivanju govora mržnje (83,9%) i isključivanju memorije AI sistema (82,5%).]

23 Na osnovu ovih rezultata mogu se izdvojiti sledeće preporuke:

1. Važan izazov medijske politike danas u Srbiji svakako predstavlja održavanje medijskog pluralizma, ne samo medijskog sadržaja već i platformi preko kojih sadržaji mogu biti dostupni građanima.

2. Nisko poverenje u informativne medije, naročito štampu i onlajn izdanja, govori o potrebi za promovisanjem kvalitetnog istraživačkog novinarstva i medija koji neguju novinarski profesionalizam, obezbeđuju kvalitetno informisanje, stvaraju pogno tlo za razvoj kritičke svesti i kritičkog mišljenja u javnoj sferi – čemu prethodi obezbeđivanje zaštite medijskih sloboda i slobode govora. Sa tim je povezana i potreba za uspostavljanjem efikasne sistemske podrške ugroženim i malim, naročito lokalnim, medijima za razliku od dosadašnje prakse sufinansiranja medijskog sadržaja preko konkursa. Način distribucije sredstava po ovim konkursima je problematizovalo nekoliko istraživanja (npr. pogledati Ranković, 20216; Matović i Purić, 20187; Nikoletić, 20208; Savić i Milošević, 20219) koja su ukazala na problem pristrasnosti pri izboru medija, nedovoljno jasnih kriterijuma i netransparentnosti konkursnog procesa. 3. Ispitanici prepoznaju važnu ulogu regulatornog tela za medije kao i potrebu da regulatorno telo kontroliše i nadzire, a po potrebi i kažnjava, medije koji emituju neprimerene sadržaje i kriše zakon i etičke kodekse. Osim jasne potrebe za regulatorom koji poseduje mehanizme, legitimitet i autoritet da odgovorno i neselektivno vrši monitoring i regulaciju medija ovo nam ukazuje na to da sami građani mogu biti važan resurs u borbi za nezavisnost i odgovoran rad REM-a osnaživanjem za korišćenje različitih mehanizama delovanja i saradnje sa regulatorom. To bi doprinelo i njihovom većem aktiviranju u medijskom i digitalnom prostoru. 4. Budući da veliki procenat ispitanika deluje uglavnom u privatnim sferama, pasivno koristeći medije i digitalne tehnologije, još je važno isticati programe koji bi povezivali jačanje kritičkog mišljenja u sprezi sa razvojem digitalnih veština koje bi građanima omogućile da sa više samopouzdanja aktivinije koriste medije bilo da je reč o zaštiti ili društvenom angažmanu,

6 Ranković, Larisa. 2021. MEDIJSKI KONKURSI MINISTARSTVA KULTURE I JAVNOG INFORMISANJA U 2020. I MEDIJSKI PLURALIZAM. BIRN & Fond za otvoreno društvo. https://nuns.rs/media/2021/11/Medijski-konkursi- Ministarstva-kulture-i-javnog-informisanja.pdf. 7 Matović, Dragana i Purić, Biljana. 2018. Analiza obrazloženja komisija za dodelu sredstava na konkursima za sufinansiranje medijskih sadržaja: nejasno i netransparentno. BIRN. https://safejournalists.net/wp- content/uploads/2019/08/Analiza-obrazlo%C5%BEenja-komisija-za-dodelu-sredstava.pdf. 8 Nikoletić, I. Danas, 28. maj 2020. Za vlast istraživanje korupcije nije u interesu javnosti. Objavljeno na Cenzolovka: https://www.cenzolovka.rs/drzava-i-mediji/za-vlast-istrazivanje-korupcije-nije-u-interesu-javnosti/. 9 Vlajović Savić, Bojana i Milošević, Marija. 2021. PRATIMO MEDIJSKI KONKURS Analiza prvog konkursa za sufinansiranje medijskih sadržaja od javnog interesa u Kragujevcu. Res Publika. https://www.gradjanske.org/wp- content/uploads/2022/01/Analiza-konkursa-2021-KG-bros%CC%8Cura.pdf. 24  odnosno učestvovanju u javnom životu. Programi treba da budu osmišljeni i realizovani tako da budu dostupni i prilagođeni različitim, pa i specifičnim kategorijama stanovništva, a naročito „ranjivim“ i onim koji najviše koriste određene vrste medija. U tom smislu, i dalje je značajno koordinisanje rada različitih aktera u oblasti medijske i informacione pismenosti (medijske i akademske zajednice, građanskih, stručnih i profesionalnih udruženja, vladinog i nevladinog sektora) kako bi različite inicijative bile uklopljene u zajednički napor da se podignu kapaciteti građana, naročito za kritičku medijsku i infromacionu pismenost. 6. Veštačka inteligencija predstavlja nove izazove za sve ispitanike, jer zahteva angažovanje svih veština medijske pismenosti (pristup, analiza i evaluacija, ali i komuniciranje, kreiranje i participacija) i savladavanje novih digitalnih veština koje bi bile u skladu sa dometima novih tehnologija. Od velikog značaja bi bilo praćenje trendova, saradnja različitih sektora (IT, obrazovni, mediji, ministarstva) kako bi se obezbedilo prilagođavanje legislative i regulatornog okvira s jedne i različitih programa obuka i upoznavanja građana sa novinama s druge strane. Važno je da programi medijske pismenosti koje uključuju AI budu planirani i koncipirani na način koji omogućava da budu dostupni različitim kategorijama stanovništva.

25