Workshop on Big Tech and Media Freedom
Author/Rapporteur: Courtney Radsch Panelists: Mira Milosevic, Khadija Patel, Courtney Radsch Chair: Julia Haas Introduction
Media freedom is inseparable from the economy and political systems in which it is embedded. It is inextricably intertwined with technology, and the role of online platforms, information intermediaries, and artificial intelligence (AI) have come to play a central role in information spaces. States thus have a responsibility to protect and promote media freedom not only by ensuring independent and pluralistic media systems and supporting quality journalism, but also in how they govern technology platforms. This should include making sure that public interest information is accessible and visible to the public when and where they seek information. The 2023 Joint Declaration on Media Freedom and Democracy recommended that human-rights centric regulation should include measures to promote a diversified and decentralized environment for online content governance and algorithmic filters such as news recommender systems where no single entity should be allowed to monopolize information flows.
Over the past decade a strong focus on online platforms, content moderation, and information intermediaries has been accompanied more recently by a focus on artificial intelligence, with generative AI raising new challenges and questions regarding training, tools, bias, fair compensation and how developments in machine learning will impact media sustainability, trust, and pluralism. Developing better understanding about how to govern technology and AI to cultivate healthy information spaces where public interest media can flourish is thus essential.
In response to these challenges, the OSCE convened a workshop and panel on Big Tech and media freedom at the 2024 World Congress of the International Press Institute (IPI) in Sarajevo. Participants examined the role of Big Tech platforms (or Very Large Online Platforms as termed by the EU) which monopolize internet search services, social media, and digital advertising. These corporate gatekeepers control much of the modern public sphere by amplifying, banning, and manipulating the flow of information and ideas through their algorithmic intermediation and by frequently changing terms of service, and they can unilaterally shape the visibility and viability of journalism around the world. This workshop and report build on previous work done by the OSCE Representative on Freedom of the Media (RFoM), in particular, the project putting a Spotlight on Artificial Intelligence and Freedom of Expression (SAIFE) that has explored the challenges that AI poses to media
Introduction 2 pluralism, access to information, and content governance and which provide guidance on how to better protect freedom of speech in a digital context. The workshop followed on a report Addressing the Impact of Artificial Intelligence on Media Pluralism and Public Interest Information by David Kaye, which was also based on a workshop held at the 2023 IPI World Congress.
The increasing use of generative artificial intelligence chatbots, content generation, and search capabilities since the launch of ChatGPT in late 2022 has significantly impacted the digital landscape, wider economy, and human rights. This technological advancement has reinforced many of the critical challenges posed by Big Tech to media freedom and sustainability in the digital age — from their digital advertising and search monopolies to their control over content moderation and access to audiences — that have yet to be addressed, underscoring the need for concerted action by States, civil society and the private sector to pursue more impactful interventions for the protection of human rights. Ensuring healthy information spaces is imperative for democracy, peace and security, and leveraging technology to enhance the availability and accessibility of quality media and public interest information online through improved governance is crucial, especially as AI is rapidly adopted and integrated into the economy and information systems. The urgency for actionable solutions and coordinated policies is heightened by the numerous elections occurring throughout 2024 in the US, the European Union (EU), and more than 80 jurisdictions worldwide.
Journalist Will Oremus observed in 2017 that the tech industry got labeled “Big” out of loathing and fear, in preparation for battle, not out of respect. The terrain for this battle has expanded to include generative artificial intelligence. This technology is exponentially more powerful than even the internet or electricity according to some of its most ardent corporate supporters. The media, civil society, academia and the public more broadly along with States and international organizations continue to find that Big Tech does not do enough to meaningfully address a range of freedom of expression, harassment, privacy and disinformation issues online and concerns that AI will worsen these issues is nearly unanimous. Normative efforts to convince dominant technology platforms to voluntarily abide by human rights standards thus need to be complemented by a wider range of policies that address the political economy of media freedom. This requires an understanding of how political, economic and social forces influence the structure, control, and function of media systems and how these affect the extent to which media can operate independently.
Introduction 3 Recent advances in AI have intensified concerns about the concentration of power in the hands of a few corporations that control the platforms for public discourse and information access — both of which are fundamental to human rights and consequently for democracy, peace and security. There is widespread agreement on the inadequacy of self-regulation of online platforms and Big Tech, underscoring the need for better legal regulatory design and improved multi-stakeholder coordination. Expanding the frameworks and tools used by States and civil society is essential to uphold media freedom and human rights, given that these rights are increasingly shaped by the way markets, data, and transnational technology are governed. This also means that it is imperative to improve media freedom literacy, especially for the public to “understand the ethical and legal implications of the media in today’s digital information ecosystem and communicate effectively, including by creating public interest content,” as the OSCE Representative on Freedom of the Media (RFoM) has outlined.
Introduction 4 Impact of Big Tech and Emerging Technologies: Reinforcing Media Dependencies on Tech Monopolies that Risk Undermining Democracy
Power discrepancies and media dependencies are at the center of the debate over the impact of Big Tech and emerging technologies, particularly generative AI, on media freedom. State authorities must focus on rebalancing and mitigating these dynamics to ensure independent public interest journalism and support public interest media. Civil society must be included and supported to help shape the laws, regulations, and policies used by States to govern technology, and can help monitor and assess implementation and outcomes, provided that they are granted access to the data they need to do so.
The dominance of transnational Big Tech platforms in national information communication ecosystems has significant implications for media monetization, audience reach, and influence. Google Search rankings and Facebook’s newsfeed algorithms can make or break a media outlet or journalist, while recommendation algorithms like those on YouTube, X (formerly Twitter), or TikTok determine visibility and monetization of journalism around the world. US tech corporations Google (which owns YouTube and dominates more than 90 percent of the Search market in most countries) and Meta (which owns Facebook, Instagram and WhatsApp) have monopolized digital advertising around the world, garnering the lions’ share revenue away from publishers while controlling access to audiences and key aspects of the digital publishing process.
During the RFoM workshop, participants described experiences that echoed the findings of studies about the impacts of content governance on journalistic content. The myriad ways that unilateral content moderation decisions reduce the visibility of news based on companies’ often changing policy priorities leave media scrambling to adapt. This risks undermining independent editorial decision making. Another worrying trend is the use of corporate content moderation systems by States — where governments seek to influence or censor information online and are assisted in doing so by Big Tech that otherwise claims to uphold free speech.
Impact of Big Tech and Emerging Technologies: Reinforcing Media Dependencies on Tech Monopolies that Risk Undermining Democracy 5 Moreover, decisions by social media and search platforms to temporarily or permanently block, shadowban, or de-amplify, demonetize or otherwise algorithmically act on media content or accounts (as well as State accounts and others) are rarely provided with adequate information about why these decisions were taken or how to effectively appeal. Time and time again, publishers and journalists describe how a handful of tech intermediaries enjoy unaccountable power to shape the visibility and viability of media, having impacts that are not in line with human rights standards and obligations regarding the free flow of information. Generative AI is poised to exacerbate this dynamic as companies shift to generative search and enable content generation at scale, which will diminish referral traffic and thus the digital advertising revenue such traffic generates, further exacerbating the challenges posed by low-quality and manipulative information.
For the public good of democracy, peace and security, content governance frameworks also must consider the imperative to provide prioritization or protection for quality journalism and public interest content. Participants observed that other sectors, such as online businesses and digital marketplaces, are better protected and the visibility and viability of their online presence enjoys stronger legal protection than that of publishers. Participants observed that COVID-19 policies proved the potential of platforms to prioritize authoritative information, and were concerned about the decisions of Google, Meta, and X to dismantle many of their efforts to combat disinformation in this year of elections. The challenges of disinformation and addressing it is further exacerbated by an increasing politicization of disinformation research, with political interference and attacks on independent research further complicating efforts to objectively study and counter disinformation.
Furthermore, the unresponsiveness of these tech companies to journalists, publishers and even governments is particularly acute in countries without corporate representatives, in small markets, and generally outside of the US and large EU countries. Even “Global North” countries struggle to govern these platforms, and participants observed that State obligations should extend to include media freedom in other States. The way that States govern technology and AI companies in their country can impact media freedom in third countries. The inability of publishers to engage with platforms, an insufficient representation in countries of operation, and the inadequacy of self-regulation highlight the necessity for legal and regulatory frameworks to ensure accountability and transparency, rebalance power dynamics, and govern the public sphere. Generative AI, moreover, is replicating and reinforcing these issues through the unauthorized, uncompensated and illegitimate use of journalism content to develop, train, and run AI systems, resulting in unsustainable business models for journalism that undermine media freedom.
Impact of Big Tech and Emerging Technologies: Reinforcing Media Dependencies on Tech Monopolies that Risk Undermining Democracy 6 Participants pointed to the systematic weakening of new media over the past 15 years which has left them feeling vulnerable as yet another disruptive technology propelled by Big Tech threatens to exacerbate these vulnerabilities.
Investigative media, which, according to participants, should be an important partner for Big Tech, face particular challenges given the expense required to do this type of journalism and the difficulty in recouping an outlet’s investment. The ease of replicating and summarizing resource-intensive investigative journalism in ways that do not provide a financial return to the media outlet coupled with the pernicious impacts of content moderation interventions that reduce the visibility or circulation of this critical work further harm this increasingly precarious form of watchdog journalism. Generally, the ability of AI applications to replicate or even regurgitate original reporting without adequately citing or generating traffic is a growing concern. Chatbots and information retrieval applications, like ChatGPT and Perplexity.ai, generate content based on predications drawn from its training data. Dozens of copyright lawsuits and innumerable examples of generative AI agents spitting out what would in other domains be considered plagiarized articles and photos have shown how outputs very often resemble the books, journalism, and art that it has in its training data, even repeating huge chunks of virtually identical content.
Although several antitrust investigations and cases are underway in Washington D.C. and Brussels, most jurisdictions feel constrained in their ability to govern some of the most powerful corporations shaping their information spaces and media freedom. Yet there is widespread agreement that AI, including the challenges of disinformation, deepfakes, and datafication are deepening a variety of crises in many parts of the world.
Recent developments have dismantled the myth that media freedom can be achieved solely by reducing State intervention, instead it has highlighted the need for State regulation focused on human rights and public interest. Big Tech corporations have repeatedly decided to deprioritize news on their platforms, with examples of blocking journalism in jurisdictions like Australia, Canada, and California, while simultaneously rolling out disruptive new generative AI products that have been trained on, use, and rely on journalism content without permission or compensation. At the same time, the risk of political interference and State capture without adequate independence and accountability must be considered and may limit the availability of some approaches. It underscores the importance of governing Big Tech in established democracies and jurisdictions such as the EU, UK, and especially the US — where most Big Tech are located — with policies centering on human rights and rule of law. “There’s a responsibility on organizations and [on] more important markets to cause regulations that can fundamentally change the business model of the platforms,” noted a workshop participant.
Impact of Big Tech and Emerging Technologies: Reinforcing Media Dependencies on Tech Monopolies that Risk Undermining Democracy 7 The EU Digital Services Act (DSA), the EU Digital Markets Act (DMA), the European Media Freedom Act (EMFA), and the EU’s AI Act are viewed as providing regulators and tech companies operating in EU countries with a set of tools that are seen as critical to improving media freedom and addressing some of the pain points with Big Tech with respect to treatment of journalistic content and data transparency. These pan-European laws are about both the single market as well as protecting fundamental rights, illustrating the importance of market mechanisms as enablers of human rights.
Workshop participants agreed on the imperative that States use a more expansive set of legal regulatory tools and coordinated policies at national, regional, and global levels. Opportunities for regulators to meet and coordinate approaches are important for learning and effective implementation, particularly considering the new EU policy frameworks.
There was widespread interest in exploring how industrial policy, public utility and universal access frameworks, and antitrust could be incorporated as part of the media freedom agenda given the impact that markets and powerful economic actors like Big Tech companies have on human rights. However, the capacity of civil society and media to leverage industrial policies is currently limited. “We are used to arguments based on human rights, freedom of expression, conventions, international documents, but when it comes to industrial policy, we often don’t know the language, we don’t know the rules,” observed a participant, echoing the sentiment of others. Civil society also lacks expertise and capacity in AI, as do many State authorities, and many are dependent on some funding from Big Tech corporations, which can put them in a difficult position. Despite more than a decade of human rights-focused interventions and voluntary frameworks, many of the online harms and platform policies that undermine media freedom remain, indicating the need to consider additional approaches and institutions that could be used. By leveraging competition policy, antitrust, and industrial policies to address Big Tech’s influence on media freedom nationally, States could also help improve media freedom beyond their borders.
Impact of Big Tech and Emerging Technologies: Reinforcing Media Dependencies on Tech Monopolies that Risk Undermining Democracy 8 Ensuring a Sustainable Public Interest Media Ecosystem
Regulations are critical to structuring the market and information environment, to ensure that freedom of expression, media freedom, and access to information flourish, while at the same time States must adhere to human rights standards that ensure media independence and public interest orientation. Regulation that requires simple tweaking of platform rules has proven ineffective. The interplay of international, regional, and national governance is essential, with the EU having taken a significant role in setting guidelines and standards that impact both Member States as well as other countries around the world. Workshop participants sensed that States with the power to regulate Big Tech must do more to address these corporations’ impacts on media within and beyond national borders, emphasizing the responsibility of major markets to adopt regulations that would address platform business models, such as unbundling, non-discrimination requirements, and/or rebalancing power dynamics between media and tech platforms.
Participants agreed that expertise is needed to address the freedom of expression and human rights threats from Big Tech, to prove harm to consumers, and to demonstrate competition failures, all of which are interrelated. In order to use a wide range of strategies “you need to prove the harm for consumers and you need to prove the failure of the competition and also suggest new solutions as to how this can be systemically resolved to have a healthy information environment,” advised a participant, noting that this may require new competencies of civil society organizations. It was also noted that different strategies and advocacy are needed if a state institution with regulatory powers already exists, but might lack expertise or be politicized, versus if it doesn’t exist at all and must be created and funded.
The media communities that have engaged with competition authorities to address market imbalances that enable Big Tech to undermine media freedom, as was the case in Australia, Canada, France and elsewhere, have looked to news media bargaining codes and enforced copyright as part of these efforts. While news media bargaining codes have been criticized for their potential to favor large media over smaller players , empirical evidence from Australia shows that journalist hirings increased and that smaller players were able to benefit, though challenges remain, including the potential for relevant companies to censor or remove news from their platforms rather than be subject to the law. In addition to public support and public-private
Ensuring a Sustainable Public Interest Media Ecosystem 9 partnerships, bargaining codes can empower the news industry by enabling collective negotiations and creating a regulatory framework that requires Big Tech to provide access to relevant data. They could be applied to the scraping and use of content to develop and train AI systems. Transparency in the terms of news bargaining deals and defining and measuring the public value of journalism are necessary, however, especially considering existing information asymmetries and the importance of inculcating trust among publishers and with the public. Ensuring that copyright is enforced to protect media from unauthorized and uncompensated use of their content in AI systems is seen as another critical aspect of protecting media freedom, which cannot exist without economically viable business models and protection for intellectual property, particularly given the increasing popularity of generative search, chatbots, and content generation applications.
The reaction by Big Tech corporations to some of these governance efforts have contradicted the UN Guiding Principles on Business and Human Rights, including restricting access to news and information produced by journalistic entities. This type of corporate censorship and related threats to stop carrying news altogether have underscored the need for States to consider using a wider array of policies to ensure that companies doing business in their country comply with democratically enacted legislation and regulatory oversight.
To this end, workshop participants explored the idea of integrating common carriage and must carry provisions which could ensure that platforms carry all content without discrimination, as in the telecommunications sector and compatible with net neutrality principles. Must carry provisions could ensure that information intermediaries are not permitted to discriminate against local journalism, as Meta did in Canada and Australia and Google has done in California. Many countries already have must carry obligations for national and linguistic content for public media. Further study into how these principles could be integrated into online content governance frameworks and protect against platform retaliation against news outlets are needed, according to participants. “We need to articulate the public and social value of news, compelling technology companies to distribute it responsibly,” observed a workshop participant.
There was also interest in learning from post-WWII industrial policies that established public service media (PSM) as a public good. Participants discussed the importance of historical analysis, cross- regional coordination, and developing coalitions and expertise networks within the media and across different civil societies.
Ensuring a Sustainable Public Interest Media Ecosystem 10 Ensuring an independent content and platform regulatory system is critical to media freedom, as is structuring markets and governing digital platforms to support media freedom. Well-designed policies can address market imbalances, while also enticing media outlets to improve their own policies and practices to comply with requirements for inclusion in such regulations or in bargaining units. Ideally these should be set by publishers themselves rather than imposed externally, in line with self-regulatory best practices and to reduce potential political interference. Media in middle- and low-income countries face challenges pursuing some business models given the limited disposable income of the population and the limited influence of their governments on Big Tech corporations. AI is viewed as exacerbating these challenges but is also seen as potentially beneficial. “The power that the news industry has, is that the new AI platforms will depend on the news industry to produce information for those platforms that they have no intention of paying for,” a participant noted, underscoring the need to address this dynamic through public policy.
Ensuring a Sustainable Public Interest Media Ecosystem 11 Multi-Stakeholder Collaboration
Coalitions and collaboration within the journalism profession and in particular with digital rights groups and advertisers, are seen as vital to protect media freedom in the age of generative AI and Big Tech, especially given the limited resources and influence of civil society in many policy domains. Participants observed that such coalitions are becoming more common in some countries and help address expertise issues, division of labor, and agenda setting. UNESCO’s Social Media Coalition initiative in Bosnia and Herzegovina together with other countries and regions was one such example. Positive developments were also observed in countries where digital native newsrooms and digital rights groups formed national coalitions to learn, strategize, and in the best cases, coordinate and advocate for policies that promote media freedom. Learning from different jurisdictions and historical examples of public interest regulation is important and more technical expertise is needed to understand how generative AI will affect journalism and the political economy of media freedom.
Strategic coalition building and engagement with small/native platforms, advertisers, and other impacted industries are viewed as necessary to engage in and divide policy work, which is likely to lead to more effective regulation of both Big Tech and AI. “Are there other industries that we can learn from and partner with so that it becomes not just media versus tech?” asked one participant, underscoring the need to work collaboratively to develop broader rights-respecting oversight. There was interest in engaging more with the private sector, for example with advertisers who lose out as well in AdTech systems that are opaque and inefficient and will also be impacted by the transition to generative search and chatbots.
Participants were particularly interested in understanding what can be learned from the music industry and other industries impacted by generative AI. The music industry was able to adapt to the digital age in part because of public policy and copyright enforcement, developments in digital rights management and tracking technologies, and because of representative industry groups that could collect and negotiate on behalf of individual musicians, performers and composers. It developed a royalty system and alternatives to the illegal music sharing sites like Napster, which violated copyright and undermined the business model of the music industry. Understanding what happened to individual artists, composers, and performers versus corporate industry interests is also important, and could be seen as akin to the journalist-publisher dynamic in the media sector.
Multi-Stakeholder Collaboration 12 Involving a full range of relevant policy actors in AI governance, however, will require more support to civil society as well as improved media literacy among officials and regulators. Expertise and collaboration are insufficient, however, when there is no independent competition body or when regulators are captured by the State or Big Tech. A recent good practice example of multi- stakeholder policy development that participants identified is the European Media Freedom Act (EMFA). The negotiations preceding this Act exemplified coordination and collaboration on controversial issues like media privileges and led to a key piece of legislation that includes significant protections and rights for journalism while imposing obligations on tech platforms. Similarly, State consultations ensured civil society were included in shaping the DSA and AI Act, although participants observed that this kind of policy work requires dedicated resources and attention but is more difficult to fund.
Implementing fair competition policies that address market failures and imbalances which affect information ecosystems, including digital advertising, licensing and copyright, and collective bargaining, and AI, were seen as promising avenues to complement traditional human rights approaches. Participants underscored the need for privacy-respecting data transparency laws and access to commercial information needed to make competition and valuation assessments.
Workshop participants emphasized the critical need for leveraging a wider range of policy and regulatory frameworks, improving multi- stakeholder collaboration, and strategic coalition building to address the challenges posed by Big Tech and emerging AI technologies to ensure a sustainable and healthy information ecosystem.
Multi-Stakeholder Collaboration 13 Recommendations
For the OSCE Representative on Freedom of the Media • Advocate and develop guidance for stronger regulatory frameworks that ensure media freedom and access to information of public interest in the age of generative AI. • Develop safeguards to help ensure that regulatory frameworks for Big Tech and generative AI are not abused to limit media freedom and clamp down on dissenting information. • Support learning and analysis of approaches from other industries, like the music industry, and evaluate the role of market power and extra- jurisdictional actors. • Collaborate with other international organizations to convene cross- sector and regulatory learning and coordination at national and regional levels.
For States • Adopt multistakeholder approaches to policy development and ensure that independent civil society organizations, journalists, news media organizations and other relevant stakeholders are consulted and included in these processes. • Promote regulatory collaboration and learning across jurisdictions • Ensure that regulatory bodies, including competition authorities, with oversight or influence over media pluralism and diversity are independent and sufficiently funded. • Improve media freedom literacy among regulatory bodies and competition authorities and leverage a broader range of legal regulatory strategies to protect and promote media freedom. • Ensure that all regulatory and legal authorities (incl. competition and antitrust) understand how human rights obligations pertain to their work, for example through training and instruction, and provide opportunities for them to learn from each other and coordinate where needed and appropriate. • Explore how market failures affect information spaces and media freedom. • Consider leveraging industrial policy to promote independent public interest media and healthy information spaces, including support to public service media. • Ensure that the relevant regulatory bodies are independent from political interference and have the resources and authority to effectively govern Big Tech companies operating in their jurisdictions.
Recommendations 14 For Civil Society • Participate in consultations related to the development of laws, regulations, and policies used by States to govern technology and AI corporations and their platforms. • Monitor and assess implementation and outcomes of media, tech and AI regulation and encourage coordination and collaboration across sectors to address challenges posed by Big Tech and emerging technologies. • Improve familiarity with a wider range of market-based approaches and legal regulatory bodes, including competition, antitrust, and industrial policy that can help enhance the existing human rights traditions typically relied upon to promote and protect media freedom. • Develop digital media coalitions and expertise networks to share information, strategies, and collaboration around AI policy.
For Donors • Ensure funding is available for policy work, including lengthy, often mundane processes and coordination among civil society.
Recommendations 15