2. Implementing policies to enhance the transparency, accountability, and plurality of information sources

2.2.1. Moving beyond self-regulation and clarifying the role and strategies of state-led policies

Self-regulation, which takes place when a group of firms or individuals exert control over their own membership and behaviour, has to date been the predominant approach taken in setting standards for online platforms. In information integrity, self-regulation refers to voluntary compliance to codes of conduct, guidelines, and other mechanisms to address issues like content moderation, privacy, or ethical practices. Such mechanisms are widely considered to benefit from the higher levels of relevant expertise and technical knowledge of the industry – in this case the platforms themselves – which in turn helps drive greater effectiveness and efficiency.

Notably, self-regulation can incorporate diverse arrangements, from completely private to varying degrees of government engagement, including around government involvement in developing or approving draft rules (Baldwin, Cave and Lodge, 2011[5]). Self-regulation allows for flexibility and industry-specific approaches; particularly for media and journalism organisations, this approach can play an important role in building capacity of news organisations to develop quality, factual content and prevent the inadvertent spread of misinformation. Self-regulatory mechanisms, such as press councils, can also play a critical role in monitoring the abuse of laws against journalists and advocating on their behalf (Lim and Bradshaw, 2023[6]).

For example, the Santa Clara Principles present a prominent self-regulatory effort focused on issues of information integrity and transparency that is not led by governments.1 Adopted in 2018, these principles are a voluntary set of recommendations for companies that are designed to provide transparency and meaningful due process to impacted users of online platforms. The principles call for clarity of platforms’ content moderation efforts; clear notice to affected users; and a robust appeals process. The Santa Clara Principles are designed to help guide, evaluate, and compare companies’ practices and activities. Additionally, the Coalition for Content Provenance and Authenticity (C2PA) seeks to increase transparency of specific content. The C2PA was founded February 2021 by Microsoft and Adobe and included Arm, BBC, Intel, and Truepic; today, membership also includes Google, Sony, Meta, OpenAI, and several camera manufacturers, content creators, and non-governmental organisations. It addresses disinformation online by creating technical standards for certifying the source and history (or provenance) of specific content, to help verify who, how, when, and where it was created or edited, should the authors wish to include that information.2

In the Netherlands, the Ministry of the Interior and Kingdom Relations developed a Code of Conduct Transparency Online Political Advertisements in 2021 to prevent the spread of misleading information during elections, highlighting the potential involvement of the state in otherwise self-regulatory initiatives. The Code of Conduct is voluntary and open to all political parties and online platforms to help promote “transparency, privacy, security, fairness, and integrity of elections.” Notably, participation is voluntary and the code of conduct notes that it does not replace other regulatory initiatives. While compliance is not enforceable, the code provides a signaling function of illustrating good conduct (at its launch, 11 out of 13 parliamentary parties and Facebook, Google, Snapchat, and TikTok had signed) (Government of the Netherlands, 2021[7]).

And yet, without democratic oversight or reporting requirements, self-regulatory regimes may generate questions of accountability. What is more, where self-regulation operates as a voluntary mechanism, the public may end up being ill-protected by regimes that effectively control the most responsible members of a field but leave unregulated those firms that are least inclined to serve the public or consumer interest (Baldwin, Cave and Lodge, 2011[5]).

X’s announcement in May 2023 that it was withdrawing from its voluntary participation in the 2018 European Union Code of Practice on Disinformation3 points to the limitations of voluntary codes of practice and principles (Lomas, 2023[8]). The Code was the first self-regulatory instrument to which leading industry actors, including Facebook, Google, Microsoft, Mozilla, TikTok, and Twitter (now X), voluntarily agreed. X’s withdrawal was preceded by an announcement in February 2023 by the European Commission that the company’s first baseline transparency report for the Code of Practice fell short of the expectations set by the other platforms in terms of the data it provided and information on commitments to work with fact checkers (European Commission, 2023[9]), further clarifying the challenge self-regulatory tools pose in enabling transparent, consistent, and comprehensive monitoring and reporting.

Mitigating the challenges of voluntary self-regulation, co-regulatory approaches incorporate industry expertise and self-governance and can allow for governments to take over oversight, enforcement, or ratification of self-regulation mechanisms (Baldwin, Cave and Lodge, 2011[5]). For example, the European Code of Practice was updated and revised in 2022, with the aim for it to become a co-regulatory instrument and serve as a strengthened monitoring framework under the Digital Services Act’s (DSA) framework. The updated version of the Code contains 44 commitments and 128 specific measures covering issues around demonetisation and reducing financial incentives for spreaders of disinformation; increasing transparency of political advertising; reducing manipulative behaviour and fake accounts; supporting researcher access to platforms’ data; among others.

In Australia, the Code of Practice on Disinformation and Misinformation was published in February 2021 by the Digital Industry Group Inc. (DIGI). While the code is voluntary and aims to provide safeguards against harms from the spread of false and misleading content on digital platforms, the Australian Communications and Media Authority (ACMA) oversees the Code of Practice and works with DIGI and the signatories to assess signatories’ transparency reports, examine how signatories handle user complaints, and encourage more platforms to sign up to the code (see Box ‎2.1).

The limitations posed by existing self- and co-regulatory regimes increase the risk that they will not sufficiently mitigate the threats posed by those actors that do the most to undermine information integrity in democracies, as well as by those who merely do not wish to engage. Such risks point to the importance of government involvement in designing, enforcing, and updating regulatory responses, where relevant and appropriate. While designing policies that protect and promote freedom of expression and active, well-informed democratic debate require engagement with civil society and private sector actors, responses cannot be left to them alone. This said, these self-regulatory efforts have had value over the years in fostering dialogue between governments and platforms on the issues at stake and helping to identify the various policy options at hand. These experiences provide an important basis on which to build.

2.2.2. Policy levers to encourage accountability and transparency

As noted in the introduction, given risks to freedom of expression that content-specific policies raise, responses should largely focus on clarifying the responsibilities online platforms have regarding their role as essential actors in the information space. In this respect, governments should ensure a clear and predictable legal framework, where the rules are clear to avoid incentivising privatised censorship (Council of Europe, 2021[11]).

Furthermore, the largely opaque operations of major tech companies prevent understanding of the role of online platforms in shaping the information environment and the actions they have taken to mitigate harmful behaviours (Lai, Shiffman and Wanless, 2023[4]). Strategies focused on increasing transparency can help build understanding around how online platforms operate and can help ensure that online platforms’ rules and implementation are clear, predictable, and proportionate. Because of the information asymmetry between online platforms and governments about how content is spread and what interventions work, transparency is also an important tool in helping governments and independent researchers better understand the information space, which in turn will help monitor the impact and effectiveness of responses and inform policymaking (OECD, 2022[3]). This opportunity speaks to the broader need to enhance measurement capabilities of relevant policy actions in this space.

Online platforms do not generally have an incentive to share information with researchers, regulators, or the public on policies, processes, algorithms, or content flows primarily due to cost, privacy, and competition concerns. By making information more accessible and accurate, policies may help ensure information is provided to facilitate better understanding of the information space and the actors therein and allow for independent verification of platform claims. Risks in other industries provide meaningful examples in this respect: until governments required its disclosure, accurate information was unavailable to the public in markets as diverse as the nicotine content of cigarettes, fuel economy for cars, or food safety (Baldwin, Cave and Lodge, 2011[5]).

Several laws have recently been implemented or discussed that focus on a wide range of transparency issues. The European Union’s Digital Services Act (DSA), the UK Online Safety Act, as well as draft U.S. legislation, such as the Digital Services Oversight and Safety Act, and the Platform Accountability and Consumer Transparency Act, all reflect growing demands for greater platform transparency (Lai, Shiffman and Wanless, 2023[4]). Government regulation to promote transparency and accountability can also build on existing or similar self-regulatory efforts, as seen in the European Union, where the voluntary Code of Practice on Disinformation is now embedded in the DSA.

Greater transparency is only part of the solution for the problem of information manipulation on social platforms. Artificial amplification of content, for example via social media bots disguised as human users, can distort conversations online by boosting the apparent popularity of certain messages and accounts. This artificial amplification can be particularly harmful during elections, natural disasters, or other crisis situations.

Governments are increasingly identifying policy responses to improve the authenticity of the information space online. For example, in 2018, California introduced the Bolstering Online Transparency Act (BOT Act), which prohibits online bots from hiding their identities to appear as a human user (State of California, 2018[12]). In 2023, the Lithuanian Parliament began discussions regarding amendments to the Law on Public Information and the Criminal Code, which could give Lithuanian government the right to order social platforms and other information providers to “remove artificially increased numbers of page views, comments, shares, likes, followers, and/or subscribers of content within eight hours, or to withdraw the possibility of access to this content.” The discussions also included the potential for criminal sanctions and imprisonment for the artificial dissemination of content on platforms.4

Across policy responses, consideration should be given to their potential impact on competition. Larger online platforms are better equipped to navigate more onerous liability and transparency rules (such as through buying or developing filtering technologies and complying with deadlines for removing and reporting on content) (Council of Europe, 2021[11]). Specifically, it may be useful to vary the extent and burden of mandated transparency relative to a platform’s size, so that compliance does not become a barrier to entry. For example, the DSA imposes additional requirements for Very Large Online Platforms (VLOPs) and Very Large Online Search Engines (VLOSEs) related to identifying and auditing systemic risk, enhanced transparency reporting on content moderation, advertising transparency, and access to data about content shared on the platforms (European Union, 2022[13]).

Ultimately, it is important to outline the specific objectives, values, and aims that increased transparency requirements are seeking to achieve, as there are several trade-offs and considerations that governments should bear in mind. Regulations in this space, where relevant, should be guided by proportionality, as designing and delivering regulations in a proportional way is an essential approach to improving efficiency, strengthening effectiveness, and avoiding unnecessary administrative burden (OECD, 2021[14]). Policy responses focused on platforms should be used as a mechanism by which governments – and the public more widely – can better comprehend and respond to the behaviours and business models of key actors whose technology dominates that space, understand and mitigate specific risks, and build knowledge of the information environment more widely.

To that end, policies encouraging accountability and transparency for online platforms and services may apply to a wide range of topics, including:

• the role of online intermediary liability protection in balancing platforms’ roles and responsibilities,

• transparency around moderation policies and policy development, risk assessment and management processes, and algorithms, to provide valuable comparative information on how online platforms operate, and

• increased transparency of online platform behaviour and content data to build understanding of the information space.

2.3.1. Information integrity requires a focus on journalist safety, transparency, and preventing media capture

Ensuring freedom of opinion and expression requires uncensored and unhindered access to the press and other media. To that end, establishing mechanisms to protect journalists and systematically investigating, monitoring, and providing access to justice for threats and attacks are also essential to ensuring journalists have the freedom to participate fully in the democratic process (OECD, 2022[20]). For example, the modified Luxembourg Criminal Code (Loi du 7 août 2023 portant modification du Code pénal) includes new penalties for attacks against journalists during demonstrations. In addition, persons who threaten individuals can be subject to imprisonment, with an aggravating factor if the target is a journalist. The code also specifies that the disclosure of private and professional information (‘doxxing’) can lead to criminal liability for the perpetrator, with again an aggravating factor if the target is a journalist (Grand Duchy of Luxembourg, 2023[42]).

Beyond a focus specifically on journalists, a related avenue to help prevent interference is to mitigate media capture and promote editorial independence. Media capture refers to situations where individuals or groups exert significant control over media organisations in a way that influences content and coverage. In these contexts, the media’s ability to serve its democratic role as a “watchdog” is compromised (Nelson, 2017[43]). The risk of capture of a media outlet by political or private interests increases as the sector becomes more concentrated (Government of France, 2022[44]), where media ownership is consolidated in the hands of a few entities or individuals. These owners can in turn promote one-sided views that can lead to polarisation and impede balanced and diverse democratic debate (OECD, 2022[20]).

Policies can play a role first in maintaining a diverse and pluralistic market for traditional media by limiting market concentration in the sector. For example, policies can take the form of control on cross-media ownership (i.e. controls on joint ownership of broadcast channels in the same geographic region). Indeed, laws designed to prevent concentration proactively often form the main pillar of a state’s efforts to guarantee media diversity and prevent concentration of opinion in the media sector (European Audiovisual Observatory, 2016[45]) (Nelson, 2017[43]). Notably, the EU Media Pluralism Monitor12 is a tool that measures the state of media pluralism across 34 countries and makes recommendations for policy action.

Promoting diversity of media ownership through anti-trust and fair competition rules involves a range of considerations. A report by the French government recommended assessing the impact of transactions on pluralism on a case-by-case basis, using an analysis based on qualitative indicators (promoting diversity of content, independence of information) and quantitative indicators (audience, coverage, economic viability of the operators) (Government of France, 2022[44]). This approach is similar to that taken in the United Kingdom. The 2003 Communications Act outlines public interest considerations for broadcasting and cross-media mergers, including that there be a “sufficient plurality of persons with control of the media enterprises”; that there be a wide range of broadcasting available that is both high quality and calculated to appeal to a wide variety of tastes and interests; and for media to have a genuine commitment to the accuracy and impartiality standards laid out elsewhere in the statute (Government of the UK, 2003[46]). Norway introduced an obligation for media enterprises and owners to provide information about ownership interests to the Norwegian Media Authority in order to create greater transparency, awareness, and knowledge of ownership interests in Norwegian media (Government of Norway, 2016[47]).

Second, policies that reinforce transparency in countries’ media markets can play an important role in ensuring media independence from political and commercial interests and freedom from foreign or domestic political influence. Opaque ownership makes it difficult to identify underlying bias, potentially further undermining trust in the news media. Transparency is therefore a necessary – but not sufficient – policy response to reinforcing media plurality and increasing trust in the media sector (Craufurd Smith, Klimkiewicz and Ostling, 2021[48]).

Notably, the European Court of Human Rights has recognised a positive obligation on States that are parties to the European Convention on Human Rights to “put in place an appropriate legislative and administrative framework to guarantee effective [media] pluralism” and that such plurality cannot be fully effective without clear information. To that end, it recognised the value of media transparency and independence to democracy, specifically in the interests of individuals in having access to information “on all matters of public interest” and the ability of the media to perform their “vital role of ‘public watchdog’” (European Court of Human Rights, 2001[49]). In addition, the 2018 Council of Europe Recommendation on media pluralism and transparency of media ownership notes that media freedom and pluralism are “crucial corollaries of the right to freedom of expression…and… are central to the functioning of a democratic society as they help to ensure the availability and accessibility of diverse information and views, on the basis of which individuals can form and express their opinions and exchange information and ideas” (Council of Europe, 2018[50]).

Requirements include transparency around media ownership, for example, by mandating full disclosure of owners, the size of the shareholdings, and their other economic and political interests. Ownership should refer to the “beneficial owner,” or the “natural person(s) who ultimately owns[...]and/or exercises ultimate effective control (FATF, 2023[51]).” The information provided should also “identify the natural person(s) who are the beneficial owner(s), and the means and mechanisms through ownership, control or other means (FATF, 2023[51]).” Such information can provide policymakers, regulators, and the public with the relevant data needed to develop, monitor, and enforce ownership limits and prevent capture (Craufurd Smith, Klimkiewicz and Ostling, 2021[48]). More can be done in this space. In Europe, for example, while most countries (24 of 31)13 require the disclosure of ownership information to public bodies, a minority (14 of 31) require disclosure to the public (Craufurd Smith, Klimkiewicz and Ostling, 2021[48]). In addition to beneficial ownership, information should also cover details of financial and other relations that could result in editorial influence and conflicts of interest, such as ownership in other industries with significant government interests, the holding of political office, and ensuring that government advertising budgets are allocated in an open and competitive way and independent of political influence (Nelson, 2017[43]).

Third, governments may also take clear positions on enforcing editorial independence. For example, Norway’s Media Liability Act seeks “to facilitate open and informed public debate by ensuring editorial independence” by mandating that publishers appoint an independent editor. Specifically, this means that the owner or company management “cannot instruct or overrule the editor on editorial issues, nor can they demand to have access to…material before it is made available to the public.”14

For its part, the proposed European Media Freedom Act seeks to protect media independence by strengthening safeguards against political interference in editorial decisions, as well as promoting transparency of media ownership and of the allocation of state advertising. It also seeks to defend media pluralism by promoting the stable funding of public service media and requiring member states to assess the impact of media market concentrations on media pluralism and editorial independence and to create a new independent European Board for Media Services, comprised of national media authorities. Importantly, it also includes safeguards against the unjustified removal of media content produced according to professional standards. This “media privilege” considers membership of press councils as one of the benchmarks for identifying reliable news media, and broadly seeks to promote media and journalism’s role in democratic discourse (European Commission, 2022[52]).

2.3.2. Governments can play an important role in supporting a diverse and independent media environment

Quality journalism is important for democracy and states should put in place effective policies to support it (Council of Europe, 2023[53]). Quality journalism, in particular quality investigative journalism, requires important financial resources. Governments can play an important role in supporting the survival and transformation of the media sector by providing various means of financial support, with safeguards around government influence on content.15 At the national level, funding can take the form of support for independent public service broadcasters; direct subsidies and competitive or selective funds for private or non-profit media; and indirect measures such as tax subsidies. Governments may also provide Official Development Assistance (ODA) as a part of their efforts to support and develop diverse and independent journalism in aid-recipient countries (Forum on Information and Democracy, 2021[54]).

2.3.3. Strengthening economic incentives to promote better functioning online information spaces

While not directly connected to counteracting disinformation, identifying economic drivers that help provide incentives to online platforms to promote information integrity is an important approach. From a consumer perspective, while online platforms have brought substantial benefits, including lower information and communication prices, greater accessibility and convenience, and access to new content and means of engagement, several concerns have been identified with respect to competition in digital markets. Notably, digital-intensive sectors have demonstrated a tendency toward greater market concentration and falling entry rates of new firms (OECD, 2019[71]; OECD, 2022[72]). This is partly a result of strong merger activity in these markets. For example, between 2001 and 2021, Google bought 258 companies; Facebook (now Meta) employed a similar practice, buying 90 companies in a period of 16 years (2005 to 2021), meaning they closed one deal every two months (Nadler and Cicilline, 2020[73]) (American Economic Liberties Project, 2021[74]). In addition, there are certain inherent characteristics of digital markets that make them prone to concentration, including the presence of network effects (the phenomenon through which the value of a product or service increases when more people use it), data feedback loops (which enable platforms that derive significant volumes of data from their large user bases to continually improve their products and services), and strong economies of scale.

Concentration may in turn have reduced competition for and availability of trustworthy sources of news (Nadler and Cicilline, 2020[73]). Moreover, with fewer options available to consumers, concentration may also reduce incentives for large online platforms to compete on quality aspects. These trends are a concern because evidence shows that healthy market competition helps spur innovation, as well as promote long-term growth and well-being (OECD, 2022[75]).

Several jurisdictions have implemented, or have proposed, specific policies to address competitive harms in digital markets. By encouraging new entrants and innovation, these strategies seek to spur competition between online platforms, potentially encouraging market-based incentives to healthier information spaces, though this outcome is far from certain. For example, regulations may address, as appropriate, data-related concerns, including obligations to implement data portability and interoperability measures. Enabling consumers to switch services more easily may prevent anti-competitive conduct and encourage innovation. Governments may also include issues related to the ‘gatekeeper’ status of online platforms, including measures to limit bundling and self-preferencing their own goods and services. Some regulators have also put in place additional merger requirements that increase scrutiny of attendant competition risks (OECD, 2022[75]).

The European Commission (EC), for example, has taken this approach through the Digital Markets Act (DMA). The EC has focused on creating and maintaining a level playing field for digital services; ensuring responsible behaviour of online platforms; fostering trust, transparency and ensuring fairness on online platforms; and keeping markets open by promoting a fairer business environment and encouraging new services to enter the market (OECD, 2022[3]).

The nature of the relationship between digital platforms and news publishers is complex. From the news publishers’ perspective, this relationship is characterised by a tension between the short-term operational opportunities of using digital platforms as effective channels of distribution of news content and the long-term concern to become “too dependent” on these platforms (Nielsen and Ganter, 2018[76]). From the digital platforms’ perspective, there are conflicting views as to the value of news content, particularly compared to other type of third-party content, for their businesses and revenue (OECD, 2021[77]).

In light of these dynamics, one avenue to promote competition in this space has been to put in place requirements for online platforms to remunerate news media companies for linking to content. In Australia, the news media bargaining code came into effect in March 2021. It addresses the bargaining power imbalances between specifically designated online platforms (notably, those that have a “significant bargaining power imbalance with Australian news businesses”) and publishers (Australian Competition and Consumer Commission, 2020[78]). The code requires designated digital platforms to negotiate in good faith with news businesses that have registered an intention to bargain. If an agreement about remuneration cannot be reached within three months, there is a compulsory arbitration mechanism within the framework to resolve disputes over remuneration (Australian Competition and Consumer Commission, 2020[78]). A government review found that by the end of its first year of operation, more than 30 commercial agreements had been struck between digital platforms (Google and Meta) and a range of Australian news businesses outside the code. It is unlikely these agreements would have been made without the Code (Government of Australia - The Treasury, 2022[79]).

Similarly, in July 2019, France enacted a law transposing the EU directive on copyright and related rights, including providing remuneration criteria for the use of news abstracts on online platforms (Autorité de la concurrence, 2020[80]). In April 2020, the French competition authority imposed interim measures requiring Google to negotiate in good faith with publishers and news agencies on the remuneration due to them under the law after finding that Google had likely engaged in anti-competitive conduct designed to circumvent the law (Autorité de la concurrence, 2020[80]). Furthermore, in 2023, Canada passed the Online News Act, which “aims to ensure that dominant platforms compensate news businesses when their content is made available on their services,” and creates a bargaining framework to encourage platforms to reach voluntary commercial agreements with a range of news businesses, which would proceed to mandatory bargaining and arbitration process if unsuccessful (Government of Canada, 2023[81]).

The potential downsides to this approach can be seen, however, in the restrictions to free and open linking across the internet imposed by the regulations, and the risk that online platforms remove access to professional and traditional news sources in particular jurisdictions entirely. Indeed, Meta announced that “people in Canada will no longer be able to view or share news content on Facebook and Instagram,” as the value Meta receives from allowing users to post links to news articles is less than the cost for paying the outlets for links that were previously made voluntarily (Meta, 2023[82]). Moving forward, the aim will be to continue to identify approaches that support an independent and diverse media sector, while upholding a free and open information space.

2.4.1. Risks posed by foreign information manipulation and interference

An important avenue for strengthening the information space is to recognise and respond to threats of foreign malign interference. If done transparently through official channels, foreign influence is legal and can contribute to democratic debates. Risks to democratic processes arise, however, from efforts by foreign agents to interfere in democratic processes and information spaces in ways that undermine decision-making, reduce trust in democratic systems, increase polarisation, and that hide the actors’ activities and intent.

While a single, universally accepted definition of foreign interference does not yet exist, the concept broadly refers to efforts by foreign actors to interfere illegitimately in decision-making processes of a target country. It encompasses actions both by state and non-state actors, as well as their proxies. Foreign interference is also marked by the co-ordination of activities and the malign nature of actions that seek to negatively impact values, procedures, and political processes. While all governments seek to influence deliberations on issues of importance to them as part of their foreign policy toolbox, globalisation and digitalisation have amplified the challenge of foreign interference and made it much more of a civilian concern, with open democracies being more fragile to foreign interference than more closed systems. Several governance loopholes can be addressed in this regard to make democracies more resilient to foreign interference.

In the information space, foreign information manipulation and interference (FIMI) seeks to shape public opinion and discourse, often with the aim of strengthening parallel interference efforts (see Box ‎2.6 for definitions). Foreign malign actors often seek to exploit global information flows to gain influence, affecting countries globally, contributing to democratic backsliding, and threating political instability and violent conflict through disinformation campaigns (Office of the Director of National Intelligence, 2023[84]).

Domestic or foreign actors may spread disinformation as part of a foreign malign influence operation. Domestic actors can act as the witting or unwitting proxies of foreign malign actors, motivated by political, economic, social, or monetary gains. A key objective of FIMI actors is to destabilise society and government within the target state and confuse public debate around key issues, with disinformation often to be designed to be spread through domestic discussion and online. One tactic used to achieve this is exacerbating existing political and social fissures. This approach allows foreign actors to achieve more effective and seemingly authentic outreach, to save resources, and to hide the origins of the interference activities.

FIMI operations often seek to influence specific domestic and foreign policy decisions of target states, sow divisions in societies, denigrate democratic values, processes and institutions, and rally support for the policies of the perpetrating state (EEAS, 2023[89]). Foreign and malign information initiatives also seek to weaken target states by targeting foreign policy interests, as well as reducing the population’s trust in government institutions, widening political cleavages and societal polarisation, and undermining democratic resilience (U.S. Department of State, 2020[90]) (OECD, 2022[91]).

Foreign state actors have also used a wide range of channels, tools, and practices to create and spread disinformation through potentially vast networks consisting of official, proxy, and unattributed communication channels, including state-backed media, global television networks, fake social media accounts and fake news websites. One avenue is via state-owned and controlled media of authoritarian states, such as Sputnik, RT, and TASS in Russia, and Xinhua and CCTV in China. The importance of these channels can be seen in Russia, for example, where government spending on “mass media” for the first quarter of 2022 was 322% higher than for the same period in 2021, reaching 17.4 billion roubles (roughly EUR 215 million). Almost 70% of Russia’s spending on mass media in Q1 2022 was spent in March, immediately after Russia’s invasion of Ukraine (The Moscow Times, 2022[92]). The outlets that receive these funds, including RT and Rossiya Segodnya, which owns and operates Sputnik and RIA Novosti, are state-linked and state-owned outlets that “serve primarily as conduits for the Kremlin’s talking points and can be more accurately thought of as tools of state propaganda (United States Department of State, 2022[93]) (Cadier et al., 2022[94]).

The Chinese government has expanded the distribution of content favourable to its positions through the reach of its state-owned media, purchasing foreign media outlets, and by publishing favourable content in foreign media outlets. For example, as noted in the U.S. Department of State GEC report “How the People’s Republic of China Seeks to Reshape the Global Information Environment,” Xinhua, the Chinese government’s official state news agency, maintained 181 bureaus in 142 countries and regions as of August 2021. The Chinese government has also purchased controlling stakes in media outlets in Europe, Asia, and Africa, in many cases evading media transparency rules and often shifting news and editorial coverage to more pro-Chinese positions (U.S. Department of State, 2023[95]). In addition, government-controlled media has used content-sharing agreements with foreign local media outlets to supply information products for free or at heavily subsidised prices to local media outlets, and in some cases prohibiting recipients from entering into content-sharing agreements with Western-sourced wire services. Such an approach can discretely promote pro-Chinese positions while limiting the reach of other outlets. These types of agreements – in which information provided by Chinese outlets appears in local media without attribution – risks distorting information environments and reduces the ability of citizens to make transparently informed decisions (U.S. Department of State, 2023[95]).

In response, for example, the Baltic states were the first EU countries to impose temporary bans on the broadcasting of some Russian TV channels, directly or indirectly run by the Russian state, which actively spread disinformation, propaganda and incitement to hatred. Following Russia’s war of aggression against Ukraine in 2022, the European Union introduced a Union-wide ban on broadcasting of the two Russian state-run channels, RT (Russia Today) and Sputnik. In December 2022, the European Union expanded the list of banned Russian TV channels to address the “systematic, international campaign of media manipulation and distortion of facts in order to enhance its strategy of destabilisation of its neighbouring countries, and of the Union and its Member States (Official Journal of the European Union, 2022[96]).”

Malign actors also use cyber-attacks to steal and distribute sensitive information as a more active effort to support wider disinformation campaigns. For example, prior to the 2017 French presidential election, a co-ordinated attempt to undermine Emmanuel Macron’s candidacy included the hacking and leaking two days before the second and final round of the presidential election of more than 20 000 emails stolen from the computers of campaign staff. This cyber-attack was timed to coincide with the campaign blackout period which prevents campaigning mandated by law and was co-ordinated with a disinformation campaign that in parallel spread rumours and forged documents. On X alone, a co-ordinated effort to spread related content by promoting the hashtag #MacronLeaks appeared in almost half a million tweets in twenty-four hours (Vilmer, 2019[97]). In addition to the harms caused by illegally accessing private information and the risks posed by cyber-attacks to democratic processes more widely, this campaign highlights how malign actors can use hacked governmental data, commercial secrets, and personal information to obscure and undermine public debate.

Actors can use opportunities provided by online platforms to amplify the reach of content to spread foreign information manipulation and interference campaigns. Beyond hijacking social platform accounts of elected or other public officials, malign actors pursue less overt means of artificial amplification, including by stealing accounts and creating “bot farms” to spread content. This co-ordinated exploitation of accounts post, share, and like target materials in ways that mimic – and may then develop into – actual engagement on platforms and even spread to off-line news sources.

Moving forward, generative AI technologies will provide greater opportunities for the creation and distribution of false and misleading content. Malign actors may use these rapidly evolving technologies to generate realistic looking and difficult to detect automatically fake user profiles, text, audio, and video materials, as well as to manage bot networks. To this end, foreign information manipulation and interference should be seen as part of larger efforts to undermine democratic processes. Disinformation efforts are an important national security tool for nations and nonstate actors whose goal it is to undermine democracy (Danvers, 2023[98]). Attacks against elected and public officials and candidates can directly distort the political process. Undermining citizens’ perception of the fairness, transparency, and security of the electoral process erodes trust in democratic system more widely. Maintaining information integrity is therefore a key measure to upholding the integrity of democracies.

2.4.2. Disinformation in the context of elections

When disinformation operations are strategically conducted during electoral cycles, they directly interfere with the essential core of democracy, can undermine the trust placed in the electoral process and the bodies in charge of it, discredit political opponents, increase the risk of disputed election results and sow social unrest (UNDP, 2023[103]); (International IDEA, 2024[104]). According to an IPSOS and UNESCO survey conducted in 16 countries where general elections will be held in 2024, 87% of respondents expressed concern about the impact of disinformation on upcoming elections in their country, with 47% being "very concerned" (IPSOS, UNESCO, 2023[105]). In addition, an increasingly digital environment brings new benefits and dangers in the context of elections. Technology can increase citizens’ opportunities to find useful information for their voting decisions and foster voter mobilisation. At the same time, technology-enabled solutions can also be used to influence the electorate by spreading disinformation, for instance through artificial amplification or AI-generated deepfakes and political micro-targeting.

As elections are usually planned and their dates well-known in advance, disinformation propagators can have time to organise sophisticated operations. In addition, elections can indeed be seen as an “ideal high-impact opportunity” to conduct their information influence operations (Polyakova and Fried, 2019[106]). It is important also to note that engaging in electoral interference strategies and activities do not necessarily necessitate tangible impacts on the results of the elections to have a negative impact: sometimes casting doubts on the legitimacy of the elected candidate can achieve the expected results by those interfering. In this context, it is also important to prepare a policy response, so that detection capacities can be deployed as early as possible to reduce the risk of interference. This said, it is important to highlight that no measure to tackle disinformation during elections should interfere with legitimate political debates or justify disproportionate measures restricting the free flow of information, including the blocking of content or Internet access (UNESCO, 2022[107]).

Given the role that elected officials, candidates, and political parties play in the information ecosystem, including in generating and amplifying content, and in some cases amplifying disinformation, reiterating the importance of information integrity in elections can play a key role. The Code of Conduct Transparency Online Political Advertisements developed by the Netherlands in 2021, for example, sought to prevent the spread of misleading information during elections by receiving commitments from platforms and political parties to acknowledge a responsibility in maintaining the integrity of elections and to avoid disseminating misleading content (Government of the Netherlands, 2021[7]).

A response to the threat of information manipulation in the context of elections includes the development of a wide range of government competences, often through the creation of specialised task forces, focused on justice, national security and defence, public communication, and election management which would ideally be established well ahead of the planned elections (see Box ‎2.8). Stakeholders on election frontlines, including independent electoral management bodies (EMB), political parties and candidates, journalists, and civil society organisations, need to be aware of the risks that disinformation poses to free and fair elections.

A key focus of efforts focused on countering electoral disinformation is around facilitating co-operation and co-ordination across governments to share information about relevant threats and deploy appropriate response strategies. Co-ordination enables relevant offices to work together to take appropriate action while respecting political neutrality. Governments can also focus on building the public’s long-term understanding of disinformation flows and risks and enhance preparedness ahead of elections. Civic education on a country’s electoral legal framework prevents information gaps that can be exploited by disinformation propagators. More broadly, voter education can help safeguard electoral integrity on issues such as campaign finance and advertising rules.

Government efforts in this space also enable short-term reactions to immediate information threats in the context of electoral disinformation. In recent Brazilian elections, the judiciary co-ordinated with digital platforms to facilitate engagement and compliance of court decisions around illegal content. In this way, the Brazilian government sought to establish open and agile dialogue channels during electoral periods between digital platforms and public authorities, while ensuring that any decisions taken with regards to content moderation were made in a transparent, public manner and in accordance with the country's laws.

In addition, government offices and task forces may provide timely and reliable information to citizens on how to exercise their rights, including voter registration and election day voting procedures, particularly in response to specific disinformation campaigns (International IDEA, 2023[108]).

Box ‎2.8. Ensuring information integrity during elections via special taskforces

Electoral Integrity Assurance Taskforce – Australia

In Australia, the Electoral Integrity Assurance Taskforce (EIAT), made up of agencies across federal government, was established in 2018 to provide information and advice to the Australian Electoral Commissioner on matters that may compromise the real or perceived integrity of an Australian federal election or referendum. Potential threats to electoral integrity can come in the form of cyber or physical security incidents, disinformation campaigns, and through perceived or actual interference in electoral processes. Notably, this taskforce focuses on referring information about relevant threats to the appropriate agencies in Australia and facilitates co-operation and co-ordination, enabling them to work together to take appropriate action while respecting strict political neutrality.

The Taskforce and its Board are comprised of the following agencies: Australian Electoral Commission, Department of Finance, Department of Prime Minister and Cabinet, Department of Infrastructure, Transport, Regional Development, Communications, and the Arts, Attorney-General’s Department, Department of Home Affairs, the Australian Federal Police, the Australian Signals Directorate, the Australian Transaction Reports and Analysis Centre, Department of Foreign Affairs and Trade, the Australian Security Intelligence Organisation, and the Office of National Intelligence.

The work of this task force is also complemented by AEC-led campaigns such as “Stop and Consider”, encouraging voters to think critically about the sources of electoral information they see or hear, and the AEC Disinformation Register, focusing on harmful disinformation related exclusively to the procedural aspects of conducting elections and referendums.

Electoral Justice Permanent Programme on Countering Disinformation – Brazil

Brazil’s Electoral Justice Permanent Programme on Countering Disinformation was established by the Superior Electoral Court (TSE) in August 2021, building on a similar programme established in 2019 that sought to prevent and combat the spread of mis- and disinformation about the 2020 elections.

To respond to the challenges that disinformation imposes on the integrity of elections and on democracy more widely, the Programme has adopted a “network” model, bringing together representatives from government agencies, press and fact-checking organisations, Internet providers, civil society organisations, academia, and political parties. 154 partners take part currently.

The Programme focuses on three actions: (i) Informing, which seeks to disseminate official, reliable, and quality information related to the electoral process; (ii) Empowering, which is aimed at media literacy and building societal understanding of both the threats posed by the spread of disinformation as well as civic education around the functioning of the electoral process in Brazil; and (iii) Responding, which is focused on identifying disinformation campaigns and countering its negative effects.

Critical Election Incident Public Protocol & Security and Intelligence Threats to Elections (SITE) Task Force – Canada

In anticipation of the 2019 election, Canada put in place the Plan to Protect Canada’s Democracy presenting concrete actions to safeguarding democratic institutions and processes. The Plan includes four pillars of action: enhancing citizens’ preparedness, enhancing organisational readiness, combating foreign interference, and building a healthy information ecosystem.

As a result of this Plan, Canada established a Critical Election Incident Public Protocol, which lays out a simple, clear, and impartial process by which Canadians would be notified of a threat to the integrity of a General Election.

Canada also established a Security and Intelligence Threats to Election (SITE) Task Force to identify and prevent covert, clandestine, or criminal activities from influencing or interfering with Canada's electoral process. The primary responsibilities of the Task Force are to raise awareness of foreign threats to Canada's electoral process and to prepare the government to assess and respond to those threats, including disinformation campaigns. The Task Force comprises representatives from the Communications Security Establishment, the Royal Canadian Mounted Police, Global Affairs Canada, and the Canadian Security Intelligence Service.

Source: Australian Electoral Commission (2023[109]), “Electoral Integrity Assurance Taskforce”, https://www.aec.gov.au/about_aec/electoral-integrity.htm; Government of Brazil Electoral Justice Permanent Programme on Countering Disinformation Strategic Plan 2022, https://international.tse.jus.br/en/misinformation-and-fake-news/tse-brazil-counter-disinformation-program-2022-f.pdf; Government of Canada (2021[110]), “Security and Intelligence Threats to Elections (SITE) Task Force”, https://www.canada.ca/en/democratic-institutions/services/protecting-democracy/security-task-force.html; Government of Canada (2023[111]), “Rapid Response Mechanism Canada: Global Affairs Canada”, https://www.international.gc.ca/transparency-transparence/rapid-response-mechanism-mecanisme-reponse-rapide/index.aspx?lang=eng.

2.4.3. Governments will need to respond to the changes introduced by generative AI to the information space

While risk-based regulation is increasingly used to mitigate the risks in the role that online platforms play in spreading information (including mis- and disinformation), such an approach should also respond to the role of Artificial Intelligence (AI) tools and systems, how those affect the information space, and how they are used as a disinformation tool that undermines human rights, for example by being used to silence women and members of marginalised communities participating in public life. The rapid development of advanced AI systems has indeed the potential to lead to innovations that can both benefit societies, while also posing new risks (GPAI, 2023[112]).

In the information space, generative AI20 tools may help identify inauthentic accounts or patterns, thereby helping governments improve their situational awareness around disinformation campaigns and complementing the moderation work by digital platforms. The tools may also be used to support the development of educational materials and activities, as well as to facilitate translation, summaries, and analysis, greatly facilitating and reducing the cost of these activities for public officials, journalists, and CSO actors alike (Landemore, 2023[113]).

The ability for generative AI to create and disseminate highly convincing content also raises the risk posed by rapid growth of realistic false or misleading news, articles, and visual media, posing an additional risk to people's trust in the information space, particularly online. In addition to content generation, generative AI could also help create a large amount of realistic fake profiles on online platforms, help animate networks of fake accounts, and overcome the detection capabilities recently created by governments, platforms or other stakeholders to identify co-ordinated inauthentic behaviour on platforms. By vastly reducing the cost of and language barriers to creating convincing text or visuals, and by making it increasingly difficult to distinguish between genuine and manipulated content, generative AI tools have the potential to magnify the challenges already introduced by online platforms. This situation may further erode the foundation of trust that individuals place in the information they consume, leading to heightened scepticism and uncertainty.

To that end, the OECD Recommendation on Artificial Intelligence calls for AI actors to commit to transparency and responsible disclosure regarding AI systems in order to: 1) foster a general understanding of AI systems; 2) ensure stakeholders are aware of their interactions with AI systems, including in the workplace; 3) enable those affected by an AI system to understand the outcome; and 4) enable those adversely affected by an AI system to challenge its outcome and understand the logic that served as the basis for the prediction, recommendation, or decision (OECD, 2019[114]).

Regarding the potential impact on the information space more specifically, focusing on generative AI tools (as opposed to the wider universe of AI applications and effects related to autonomous weapons, facial recognition technology, self-driving cars, and economic impacts), is a helpful framework for analysis. Policies could consider requiring that consumer-facing generative AI systems make public the training data used to build the systems, ensuring that the principles used to guide the tools are available to allow for comparison between tools and public oversight of what guardrails systems have put in place (or not, as the case may be), and watermarking of content produced (Giansiracusa, 2023[115]).

Along these lines, the proposed EU AI Act, presently under discussion, follows a risk-based approach and establishes obligations for providers and users depending on the level of risk the AI can generate. On the one hand, the EU AI Act will seek to prohibit AI systems with an “unacceptable level of risk to people’s safety”, including systems that “deploy subliminal or purposefully manipulative techniques, exploit people’s vulnerabilities or are used for social scoring (classifying people based on their social behaviour, socio-economic status, personal characteristics)” (European Parliament, 2023[116]). The act would also require the creation of risk assessment and mitigation plans and require that generative AI tools follow transparency requirements, such as disclosing what content was generated by AI. The EU AI Act would also require tools to be designed to prevent the generation of illegal content and to publish summaries of copyrighted data used for training (European Parliament, 2023[116]). The EU’s approach in this space illustrates how the application of a risk-based approach can inform other regulatory responses to technologies that play an important role in the information space beyond online and social media platforms. Similarly, governments have sought to counter the risks posed by deepfakes, audio or visual media content that seem authentic but are in fact synthetic or manipulated.

Deepfakes present a disinformation risk by presenting believable, though fake, images and audio. While synthetic media is not new, the access to technology, scale, speed, and quality of deepfakes has increased a focus on the role of policy responses. Many of the efforts to prevent risks posed by deepfakes seek to enhance transparency around the content itself and the processes followed by the systems to help validate provenance and accuracy, as well as to build on existing legal restrictions on content use. An approach focused on transparency can avoid regulatory overreach that may limit the technology’s use for protected speech, such as satire. Along those lines, the EU 2022 Strengthened Code of Practice on Disinformation commits signatories that develop or operate AI systems to report on their policies for countering prohibited manipulative practices that generate or manipulate content (such as deepfakes). In addition, many of the laws passed in US states have focused on non-consensual deepfake pornography given the clear harms caused and limited speech benefits. In this regard, nine states have enacted laws that regulate deepfakes, mostly in the context of pornography and elections influence (Poritz, 2023[117]). In 2023, furthermore, the Office of the President of the United States issued an Executive Order on Safe, Secure, and Trustworthy Artificial Intelligence, which specifically seeks in part to protect individuals from “AI-enabled fraud and deception by establishing standards and best practices for detecting AI-generated content and authenticating official content (U.S. White House, 2023[118]).”

Ultimately, by identifying, analysing, and prioritising relevant risks, taking a risk-based approach can help ensure regulation is targeted and proportional, and that it does not introduce burdensome rules with little positive impact (OECD, 2021[14]). In the information space, such an approach aims to better understand, flag, and mitigate proactively the risks posed by relevant actors and to encourage or require actors to put in place mechanisms and processes that limit the risks posed by disinformation and build trust in the information space.