The Issue

Technology has created new opportunities for journalism, making it easier for diverse voices to be heard and for journalists to connect with the public. However, these advancements have brought new challenges, including a rise in online harassment and abuse towards journalists and news organizations. Some harassment campaigns are led by individuals; others are orchestrated or endorsed by governments. 

Online harassment is generally understood as using technology to bully, threaten or aggressively target someone. It can cause journalists significant psychological and emotional distress and can be linked to real-life violence. Such abuse can also lead journalists to self-censor their work or distance themselves from their audience to avoid further abuse. One in three journalists in CNTI’s recent survey regularly face serious risks, including online abuse. Other data show that online harassment disproportionately affects women journalists and journalists of color. 

The increases in online abuse are prompting responses from legislative bodies, tech companies and civil organizations. Some countries, like Australia and the United Kingdom, have passed legislation to address online safety by defining abusive content and requiring more oversight of online service providers. In the U.S., legislation has been repeatedly introduced to modify Section 230 of the Communications Decency Act, which currently protects online platforms from liability for user-generated content. Finding the appropriate level of content moderation is difficult. While some argue for stricter content moderation to curb abuse, others worry this could threaten freedom of expression. 

To help with this issue, tech companies and civil service organizations are offering journalists tools like AI software to filter harmful content, cyber toolkits and field manuals to improve online safety. 

Despite these efforts, there is a consensus that these non-legislative actions and tools are insufficient. Journalists facing high risks report insufficient protections and feel that their newsrooms are not supportive of their needs. This has led many to deal with online abuse on their own. While methods like peer support have shown promise in mitigating harm and building resilience, further action is necessary to protect journalists from online abuse and to support those who have experienced it. 

What Makes It Complex

Any policy or standard for moderating content and mitigating online harassment must also protect the freedom of expression and free speech. Moderating online abuse is a significant challenge due to the tension between free speech protections and the difficulty in defining where expression becomes harassment. This becomes more complex with the different types of digital communication, whether the communication is public or private, and if moderation occurs before or after an individual receives it.  

Countries and regions like Australia, the United Kingdom, and the European Union have begun to formalize content moderation requirements and methods, with many of the implementation decisions resting with social media companies, though some laws provide detail about online providers’ reporting requirements and the authority of various government agencies to handle online abuse.  With declining press freedoms and increasing government censorship globally, there is concern about the potential for state actors to exploit content moderation for censorship.  The balance of decision-making power is both complicated and important: too much power in the hands of any single entity, whether a technology company, government entity or other powerful institution, could be used for information control if critical safeguards and checks on power are not in place.

Technology companies have weakened their content moderation protocols and teams, presenting novel challenges and concerns. Recent layoffs across the technology industry have significantly impacted teams dedicated to combating online abuse and disinformation. Throughout 2025, platforms, including Meta and TikTok, have continued to cut staff and programs in this space..

These staff cuts in trust and safety programs coincide with an increase in online abuse, misinformation,and harmful content. While data about the effectiveness of trust and safety programs is limited, their goal of preventing abuse remains important, though clearer articulation of these goals and how they balance freedom of expression may be needed.

While not aimed directly at journalists, these trends are taking place as legislative efforts worldwide are pushing for stricter online safety regulations. Tech companies now face substantial civil penalties for non-compliance with laws like Australia’s Online Safety Act and the EU’s Digital Services Act, highlighting the critical role trust and safety teams play in protecting individuals and journalists from online harm.

The online space has enabled and encouraged journalists to engage and make direct connections with the public which, while valuable, also increases the potential for online abuse. The digital environment allows for and encourages “reciprocal journalism” where journalists and the public have the opportunity to learn from each other, which leads to increased trust and engagement in the news. Journalists provide factual information to the public and individuals share their experiences, thoughts and reactions with journalists.

At the same time, this closer connection creates more access points for potential online abuse. The rise of online harassment has significantly influenced journalists’ perceptions of their audiences. Those subjected to high levels of abuse tend to express negative sentiments towards their readership, adopting defensive mental barriers and emotional boundaries. As a result, they may withdraw from online platforms and self-censor their content.

In this delicate balance, the promise of reciprocal journalism can be overshadowed by the threat of online abuse, undermining the symbiotic relationship between journalists and their audiences. This situation is exacerbated by the disconnect between those making the policies on online abuse and those experiencing the abuse. 

Newsrooms do not always provide support to journalists experiencing harassment, creating mistrust within newsrooms. Financial and organizational culture often prevent newsrooms from providing support, such as mental healthcare coverage, to employees. As a result, journalists feel like they must deal with online abuse alone, and many have voiced a lack of trust in their newsrooms’ capacity and ability to assist them. 

News organizations can support journalists by developing comprehensive policies and protocols about online abuse and safety. For example, a common framework for reporting, recording and reviewing online abuse can lead to understanding the efforts used to target  journalists in a news organization.

Another way to aid journalists experiencing online harassment and abuse is through peer support. While it does not prevent online abuse, encouraging journalists to connect and discuss their shared experiences can alleviate some of the detrimental effects of online abuse.

End-to-end encryption complicates the government and technology companies’ shared responsibility for online content moderation.  End-to-end encryption (E2EE) services, such as Signal, Telegram and WhatsApp, can help shield journalists from obtrusive surveillance, and many journalists frequently use these services to communicate with each other and with sensitive sources. Different countries’ policies, such as the TAKE IT DOWN ACT in the U.S.,  are made to address online abuse but threaten the integrity of E2EE services.  

E2EE moderation is also complicated because globally, laws regarding online privacy are a patchwork. Current legal frameworks make content moderation across various geographic contexts convoluted. Governments and law enforcement organizations have even pressured technology companies to provide access to private E2EE messages. To address legitimate concerns without overmoderating legal content, technology companies and governments will need to find an appropriate balance that protects individual privacy online, including E2EE, while also mitigating online abuse.

State of Research

Journalists around the world are facing a growing tide of online abuse, and research reveals important trends and impacts. While any journalist can be a target, studies show that women and journalists of color are disproportionately affected, experiencing more intense and overt harassment. This abuse has serious consequences, including psychological distress, self-censorship and leaving the industry, threatening the existence of a free, independent and diverse press. 

Research shows the need for more support from news organizations. Many journalists feel their employers expect them to be active on social media platforms but do not provide adequate policies or support to protect them from harassment. This can lead to a perception that the organization prioritizes its reputation over the well-being of its staff. 

The research also shows that these attacks are often organized campaigns of “mob censorship.” In these scenarios, political leaders or other public figures make disparaging remarks about the press, which then incites groups of followers to target individual journalists with the goal of silencing them. 

To combat this, there are several promising interventions. Newsrooms can develop clear policies and protocols to address online abuse, provide support services for their staff and freelance journalists, and educate themselves about these coordinated campaigns. Peer support groups have emerged as an effective tool, offering a sense of community and shared understanding for journalists who are experiencing similar challenges. 

Continued research is needed to explore how technology can be used to mitigate abuse, what other support systems can be implemented, and how technology companies can be more transparent and accountable.

State of Legislation

A 2023 report by the World Bank reveals that out of the 190 countries examined, 58 had cyber harassment laws, and 22 of those 58 had legislation related to cyber sexual harassment. These are significant gaps in protection from online abuse for journalists. While some of these laws can cover journalists by extension, there are very few with a specific focus on online abuse of journalists. 

Countries have also been developing comprehensive legislation regarding online safety at large. This type of legislation is broader than cyberbullying legislation in that it focuses on many online harms (e.g., exploitation, data privacy and protection, terrorism). The European Union’s Digital Services Act (passed 2022, enacted 2024) and the United Kingdom’s Online Safety Act (2023) both have provisions for reducing online harms. Australia and the United Kingdom signed a joint memorandum (2024) to collaborate on creating a safer online experience for all users. 

The United Kingdom’s Online Safety Act and the U.S.’ TAKE IT DOWN Act (2025) both impact end-to-end encryption (E2EE) of online services. The United Kingdom’s act has provisions for online service providers to monitor end-to-end encrypted services for online abuse (e.g., child exploitation, terrorism). The U.S.’ act does not provide exceptions for E2EE services in its mandates for online services to remove flagged content within 48 hours or to identify and remove known copies. These types of provisions have led to concerns about government overreach and an invasion of privacy. 

Beyond legal protections, technology companies and social media platforms are also offering tools to journalists to limit online abuse such as filters to decrease the likelihood of encountering abusive content and a streamlined process of reporting abusive content to platforms. PEN America’s report about how to fix reporting processes on online platforms provides potential revisions/updates that can increase the effectiveness of reporting abusive online content.

Notable Legislation

The Issue

In its earliest days, the rise of the global internet signaled a potential for free and open societies that would transcend borders, connect people through digital civic spaces and provide a global repository of information. The internet, as a public good, can provide access to reliable and independent news as well as exposure to diverse sources and perspectives. But amid new technologies, changing geopolitical contexts and threats of online abuse and disinformation, policymakers around the world are rethinking updates to internet governance models. 

At the core of these discussions is who controls how the global internet functions. Often, this control is negotiated between governments and technology companies that preside over online spaces. 

These political, commercial and technological pressures risk fragmenting the internet into a collection of different networks and user experiences. Two users may encounter wholly different internet experiences based on their location and/or which corporation connects them to the internet. (This process is also referred to as “splintering” or “balkanization” and the result is sometimes referred to as “the splinternet”). In fact, internet fragmentation is already occurring in some places. While practices that result in internet fragmentation — via rhetoric, technological developments and legislation — are often enacted by autocratic regimes, they are also increasingly central to internet policy debates in democratic societies.

Regulations intended to restrict and/or control online environments can involve several tactics, such as content takedowns, blocked or rerouted websites or platforms, government-ordered network disruptions and versions of the internet entirely closed off from outside providers or services.

In weighing the challenges of this issue, it is helpful to distinguish between two types of internet regulation that result in fragmentation: 

1. Those where internet fragmentation is intentionally sought out, often to assert state control over data and digital assets or to cut off public access to independent information.
2. Those where internet fragmentation is an unanticipated byproduct of governments’ (or corporations’) attempts to prevent the spread of disinformation, address legitimate online harms and/or protect residents’ data from foreign interference. 

There are crucial differences between these two approaches, but both introduce risks if they do not include safeguards that protect both user rights and a free flow of global information. The preservation of an open internet is especially critical for the global protection of an independent, competitive news media. Journalists depend on access to open digital communication channels to both document and distribute news, particularly when platforms have been infiltrated or blocked by the state. A number of organizations that analyze internet freedom have found that online censorship and manipulation are growing worldwide. These are powerful threats to press freedom and safety as well as to public access to information.

As governments consider different approaches to internet regulation, they should consider that some policy responses risk facilitating a fragmented internet while others can protect societies from it. The challenge is to ensure they protect the public’s ability to use the internet to create, share and access information as well as journalists’ ability to report and distribute it — both within and across borders. 

At its most harmful, internet fragmentation enshrines government and/or corporate power in ways that erode human rights, including those of free expression and privacy. Policymakers must pay greater attention to these effects — even when unintended — to protect and promote an open internet.

What Makes It Complex

I. Internet fragmentation does not have a singular definition. As a result, it is hard to identify and address risks with comprehensive policy.

In academic and policy circles, the concept of internet fragmentation is broad and often contested. However, experts frequently cite three dimensions of fragmentation: 

1. Technical fragmentation, where systems may be completely disconnected from the global internet (e.g., China’s Great Firewall), or there may be other challenges to interoperability, such as websites that only render correctly on certain hardware. (For example, some websites designed for larger screens are very difficult to use on mobile phones — the primary internet device in most of the world.)
2. Content or user experience fragmentation, where individuals have different online experiences. In these cases, users may have technical connectivity, but face restrictions in what they can access (e.g., government-approved sources in a Google search or a social media feed, Netflix’s content filtered by geographic controls). It is also important to note that many differences of this type are beneficial, such as language and content localization.
3. Political or governance fragmentation, where decisions are made without consensus or where national governments introduce regulations that conflict with existing processes and policies. For example, Ukraine requested in 2022 that Russian domains be disconnected from the internet as a form of sanctions, although this request was ultimately denied.

Each of these types of internet fragmentation introduces different risks, requiring different approaches to address them.

II. Bans on specific digital platforms have become a common response to concerns about online safety and disinformation, but they increase the risk of internet fragmentation and, some argue, threaten free expression.
A growing list of governments around the world have taken — or considered taking — action against TikTok and its Beijing-based parent company ByteDance. These bans respond to ongoing debates around two key issues: the potential national security risks of allowing the Chinese government to access sensitive user data and/or distribute disinformation, and rising concern about children’s online safety on digital platforms. Autocratic governments have also weaponized bans of specific platforms, and some argue that these bans inhibit fundamental rights to access and distribute information in democratic societies. Enforcing bans on specific platforms could also require technical surveillance that could put privacy and civil liberties at risk. 

Some argue that there are times when platform bans are justified. For instance, in Brazil, Telegram and X have been blocked temporarily due to failures to comply with lawful requests from the Brazilian Supreme Court. 

III. There is legitimate disagreement about the bounds of free speech, even among democratic societies.
There has long been debate over whether and to what extent freedom of expression should include harmful forms of speech like hate speech and disinformation. For example, Germany places much stricter limits on free speech than the United States. More recently, these disagreements have come to the forefront in efforts to regulate online platforms, and are at the root of tensions between the U.S. and the EU. Inconsistent definitions of acceptable speech create more opportunities for fragmentation.

IV. Internet fragmentation presents new challenges for corporations asked to comply with demands that would permit governments’ abuse of power.
At times, governments may demand (including through legislative policy) that platforms provide private data and information on their users or take down content that is otherwise protected by free speech laws in other jurisdictions. If corporations adhere to these demands, they risk enabling government abuse. If they do not, they risk breaking the law, which may lead to fines or bans. These situations have arisen in Malaysia, India, Brazil and the United States which encompass many different sociopolitical contexts, including democratic societies. It is crucial to acknowledge that this issue is multifaceted, and corporations often choose specific jurisdictions and legal frameworks for market and product reasons.

V. Protecting an open internet is not always possible in autocratic societies, but actions taken elsewhere still impact people living under these autocratic regimes.
Citizens, advocacy groups and watchdog organizations often have limited power to stop autocratic action. However, open and globally focused policy discussions raise awareness and, as such, offer a means of pushing back against abuses of power and creating greater accountability. Further, democratic governments can impact people in authoritarian regimes by taking internet governance actions — for better or for worse. For example, while the Iranian regime’s censorship and infrastructure controls form the core of internet restriction, U.S. sanctions exacerbate the problem by further limiting ordinary users’ access to global platforms and services. On the other hand, the EU’s Global Gateway promotes rights-based infrastructure in emerging economies.

VI. Consolidation of infrastructure among a few large companies may counterbalance state power — but it increases other risks.
Every layer of the internet has become consolidated among a few very large companies. There are benefits to this consolidation, notably for users’ security and privacy from the state. However, consolidation can cause problems for internet resilience: when these large companies face attacks or technical problems, the impact is much broader. Moreover, consolidation increases the reach of corporations, worsening the potential impact of state abuse by proxy and corporate-driven fragmentation. Many observers have raised major concerns about corporations’ power over both public and private information.

VII. Zero-rating programs can prevent people from accessing the complete, open internet.
Technology companies, including mobile carriers and app providers, have implemented “zero rating” programs, which provide select content or services without data charges. These programs are offered to attract customers and increase internet access, especially in less wealthy countries, but they can also fragment the internet landscape. By granting free access to certain sources of content but charging for others, these programs can become the arbiters of what content is accessible, creating a tiered internet ecosystem as they promote broader adoption. (For example, Meta offers free internet access to millions of people who would not otherwise have access — but Meta limits what they can access.) 

VIII. The contemporary internet is still not a truly universal or open network; technical and linguistic access vary considerably around the world.
Debates around public access to information and international connectivity standards must consider existing barriers to an open, global internet. According to the World Economic Forum, 2.5 billion people still lacked internet access as of September 2024. There are also large gender gaps in internet access in low- and lower-middle-income countries. Even among those who do have access, speeds vary dramatically around the world. And access gaps are not just technical: 85% of internet content is in just 10 of the more than 7,000 languages currently spoken. Even among the dominant languages, the universe of information available online differs from one language to the next. New efforts are just beginning to address this challenge.

IX. Policymakers may not have access to nuanced technical expertise in how the internet works, which can make it harder to foresee unintended consequences.
The internet is complex, and policy development requires evidence-based technical discussions in order to be effectively addressed, especially when it comes to fragmentation. Understanding the layers of the internet stack is vital to identifying technical fragmentation or changes in infrastructure that prevent systems from interoperating and the internet from functioning consistently at all endpoints. However, additional forms of expertise are also important — such as expertise in privacy, bias and disparate effects. That’s because fragmentation is not necessarily only technical in origin. Fragmentation can also be rooted in governmental policies or corporate practices that prevent use of the internet.

State of Research

To date, much of the work on internet governance has centered on “digital authoritarianism” (with a particular focus on China and Russia), as well as some research examining the European Union’s digital sovereignty agenda. We are just now starting to see a growing segment of research focused more broadly on assessing both the extent of internet fragmentation and the mechanisms that contribute to it. 

However, most of the research either provides informative case studies or focuses narrowly on the policy context. As a result, there is little systematic research on the scope and impact of shifts in internet governance. Moreover, a lack of consensus about what internet ‘freedom’ and ‘censorship’ entail also means that researchers disagree about appropriate interventions. 

As platform bans, data localization and age verification become more commonly discussed approaches to addressing online safety and disinformation, research analyzing the scope and impact of these policy debates will be particularly useful. This includes how particular policy language is adopted in one context and readopted elsewhere over time and on a global scale.

State of Legislation

The internet is now a central locus of global policymaking as governments increasingly debate how to regulate digital spaces within (and sometimes outside) their own borders. Internet governance includes a broad range of stakeholders and mechanisms including national and international policy as well as infrastructures, technical standards and company policies.

In recent years, there has been a shift towards “digital sovereignty,” or governmental capacity to control or regulate residents’ internet content and their country’s flow of data through its technological infrastructures. 

In some cases, legislation attempts to protect societies from a fragmented internet. In others, legislation risks encouraging fragmentation either intentionally or unintentionally. Such legislation includes direct attempts to develop a closed national intranet, interventions to block access to specific platforms or sources of news and information, and broader online governance laws that may introduce indirect consequences for free expression and an open internet. For example, online safety regulations in the European Union, the United Kingdom and the United States may promote protections for residents while risking user experience fragmentation by enhancing controls over what and where content is shown.

Those developing legislation to protect an open internet should consult resources and recommendations provided by global internet experts. Recommendations include limiting governments’ ability to collect or access user data, protecting encryption and VPN access, creating interoperable privacy and governance regimes, and maintaining access to services and platforms.

Notable Legislation

Archived Resources

Issue primers have been reviewed at multiple stages by more than 20 global research and industry expert partners, including CNTI advisory committee members, representing five regions. We invite you to send us research, legislation and other resources. Read more about CNTI’s issue primer and other research quality standards.

The Issue

Manipulated imagery has been around for over 150 years but it has reached a new level with “deepfakes.” The term “deepfake” originated in 2017 to describe audio and videos manipulated with the assistance of artificial intelligence (AI) to resemble a real person even when the person portrayed did not say or do what is depicted in the content. Deepfakes are a subset of “synthetic media” which include audio, images, text and video content created with the assistance of AI. There continue to be conversations about how to differentiate synthetic audiovisual content from deepfake audiovisual content. Like many topics CNTI covers, the definitional clarity of these terms remains a work in progress and is important when considering policy.  

The number of deepfakes online increased tenfold from 2022 to 2023. While some research raises questions about the degree to which harm can be directly attributed to manipulated media, there is certainly some evidence of it in countries such as Slovakia, the United Kingdom and the United States. This is especially alarming amid a record-breaking number of national elections being held in 2024 and as broader global concerns grow about threats to the overall stability of a country. A March 2022 deepfake, for example, depicted the Ukrainian President falsely ordering his country’s military to surrender. 

To date, the highest-quality, most-convincing deepfakes require a large amount of training data and a lot of time to be persuasive. Therefore, most actors intending to cause harm are likely to pursue less resource-intensive means to spread false narratives such as disinformation tactics discussed in a separate CNTI issue primer. But as technological innovations advance, deepfakes are rapidly becoming easier to make and more persuasive.

Alongside the worry about the direct impact of a deepfake is concern about a “liar’s dividend” and the sowing of further distrust in “real news,” the news media and government figures.

Responses to the growth in deepfakes are occurring on several fronts. Many digital and online platforms (e.g., Facebook, Instagram, TikTok, X and YouTube) have begun implementing disclosure policies that require advertisements using AI-created content to be labeled. Others are banning certain types of synthetic material, creating training resources to assist in combating these images and working on ways to embed content with tags to confirm authenticity. 

News organizations have also developed online training courses to assist with identifying deepfakes. However, interviews with expert fact-checkers reveal that while deepfakes are a concern, many feel that more critical threats come from text, images, audio and/or video taken out of context (i.e., “decontextualized”) as well as other forms of manipulated media such as “cheap fakes” where information is recontextualized into a false narrative. 

Methods to detect synthetic media have advanced over the last several years. These identification technologies examine the shadows, geometry, pixels and/or audio anomalies of suspected synthetic media and look for hidden watermarks to evaluate authenticity. Detection challenges have encouraged researchers and the public to experiment with new ways to accurately identify deepfakes. While these techniques will almost certainly be unable to fully erase the threats of synthetic media, they do offer steps toward establishing guardrails to protect the public’s access to authentic information. To identify how to best implement detection methods, further collaboration across various sectors (e.g., technology, communications, policy, government, etc.) is needed. 

Most countries do not have specific legislation on synthetic media, but among those that do, policies fall into two general categories: (1) banning all deepfake content that does not obtain consent from the individual(s) depicted or (2) requiring disclosure and/or labeling of deepfake content. Deepfake-related regulation is particularly complex due to many countries’ protections for freedom of speech and expression. 

What Makes It Complex

Clearly delineating what qualifies as a deepfake is difficult but critical. Governments and researchers are confronting how best to differentiate deepfakes from other types of synthetic media content. For instance, there is a question about whether the material depicted must be deceptive in nature to be classified as a “deepfake.” Other definitional considerations revolve around intent, harm and consent. For example, a 2019 synthetic video of soccer star David Beckham speaking nine languages was intended to disseminate factual information about malaria but it used deepfake technologies to make the dialogue sound authentic. The intent of this deepfake was not to be deceptive or spread harm but it is still widely considered a deepfake because of how it was made. On the flip side, altered images have been around for over 150 years – without the use of artificial intelligence – but may be intentionally deceptive, similar to AI-generated deepfakes. The spectrum of synthetic media becomes increasingly complex with the inclusion of “shallowfakes” and “cheap fakes” – forms of manipulated media that do not require advanced technological tools. Better delineating what types of content are classified broadly as synthetic media versus specifically as “deepfakes” (a subset of synthetic media) is crucial to separate the benign and beneficial uses (e.g., for education and entertainment) from harmful uses.

While developing software to detect and counter deepfakes requires strong digital infrastructure and financial resources that only certain countries have available, new labeling and disclosure tools are making methods for addressing deepfakes more accessible globally. Developing independent software for detecting and countering deepfakes is expensive, but tools to identify whether or not media sources have been manipulated are becoming available for wider use. One potential strategy is to use trained human graders in combination with pre-trained AI detection models. Researchers find these combined approaches can have advantages over using a single detection method. In response to the growing number of deepfakes, content creators and the technology industry have also begun developing ways to tag and label manipulated media. These include both direct and indirect disclosure approaches to maintain transparency and assert provenance (i.e., authenticity) as well as different types of content labeling. Watermarking is one technique that can be visible to users or embedded in media to certify its authenticity. Arriving at a global standard for this type of labeling should be a priority. The Coalition for Content Provenance and Authenticity (C2PA) is one possible standard which has received support from Adobe, Google, Intel and Microsoft among many other organizations. While technology tools and disclosure and labeling requirements greatly help to address deepfakes, they are unlikely to remove all mis- and disinformation from the news ecosystem so understanding how to mitigate threats from all sources is critical for promoting fact-based news. 

Efforts to regulate deepfake content must be compatible with laws protecting freedom of speech and expression. Governments need to determine where to draw the line between legal and illegal deepfake content. In countries that legally protect free speech, deepfakes present difficult circumstances for determining what is legal or illegal content. To a degree, sharing false statements is a protected right under freedom of speech and expression laws so banning all deepfake content is likely illegal, thus making the regulation of deepfakes particularly complex. 

Synthetic media create opportunities for journalists to protect their identity in threatening situations, but deceptive behavior runs counter to many news outlets’ codes of ethics on misrepresentation. The technological innovations brought forth by deepfakes may allow journalists and/or their sources to remain anonymous by altering their appearance and voice when working on sensitive projects. However, this goes against many outlets’ codes about journalists remaining honest and transparent during reporting as well as policies about deception. While news organizations may outline rare circumstances for journalists to protect anonymity or to engage in deceptive practices, these are only allowed for matters of public interest or personal safety. Determining when journalists can use deepfakes for their work is an important ethical consideration.

State of Research

While deepfakes are a relatively new technological innovation, much research has explored how individuals interpret false information presented in differing forms of media (e.g., text, audio and/or video). In one recent study, even in the presence of a warning that deepfake content would be encountered, nearly 80% of participants did not correctly identify the only deepfake in a series of five videos. Other work has found that up to half of respondents in a nationally representative sample cannot differentiate between manipulated and authentic videos. These findings indicate that deepfakes are difficult to counter and that correcting false information is needed to support a well-informed public. 

There are also evidence-based areas of hope. Research suggests people can be trained to better detect deepfakes. Interventions that focus on the accuracy of information and the low cost of producing consumer-grade deepfakes (which can now be done in a matter of minutes using apps and websites) show positive results in an effort to counter the negative effects of this type of content. 

To complement training individuals about detecting deepfakes, one strategy is to study how responsive individuals are to informational fact checks and labels on manipulated media: 

• Fact-checking politicians’ false statements has been shown to decrease beliefs that the statements were factual, though further research into how partisanship shapes interpretation of synthetic media is crucial. Developing digital media literacy approaches, like how to spot false information, will likely be important to help individuals recognize high-quality, fact-based news. 
• Findings suggest that tagging information as false, while beneficial, also has consequences for true, authentic information. General, broad disclosures about false information can cause viewers to discount true, accurate news. The public may grow more accustomed to looking for labels and other kinds of disclosures as a way of trusting its veracity but that remains a question for now.
• False information that is not tagged as such is found to be interpreted as more accurate than other false information that has been tagged as false. These findings suggest that labels can be effective but they must be comprehensive and cover all applicable synthetic media.
• The value of asserting provenance, or tracking the authenticity and origin of content, has also been studied. While the public may not widely grasp provenance as a concept, it has been shown to decrease trust in deceptive media when presented to individuals. Further education on the importance of provenance for a fact-based news ecosystem is needed.

Future research should continue to study (1) how individuals engage with synthetic content and (2) how persuasive the public finds this content, which is especially relevant given how realistic and life-like these media are becoming. In response to the increased presence of synthetic media, researchers should also consider what techniques – including labeling and disclosure – are most effective for mitigating the negative effects stemming from deepfake content. Understanding how to respond to and “treat” individuals who have encountered deepfake content is an important consideration that can support fact-based news endeavors. Finally, research should also examine how newsrooms will confront the proliferation of deepfake content and its potential harms.

State of Legislation

Most countries do not have any existing policies that specifically target deepfakes. The existing legislation on deepfakes grapples with how to accommodate less harmful and/or benign uses of synthetic media (e.g., art, education, comedy) while addressing a broad range of harmful uses (e.g., nonconsensual adult content, deceiving consumers). Defining the harmful, illegal uses of deepfake content is critical for effective policy. Legislation attempting to protect societies from deepfake content must decide whether to ban all AI-generated deepfake content or to set restrictions on material allowed in deepfake content and develop regulations on how it is presented to audiences (e.g., labeling). 

Another crucial consideration for public policy is how to best classify deepfake media content. Much of the legislation being debated and passed in U.S. states is about images, videos and audio. However, some pieces of legislation also address text-related synthetic media, while others focus solely on videos – even if generative AI was not used in the manipulated videos. Global standards for what types of media are classified as deepfake content might be helpful but only if they are fully inclusive of the ways people receive information and interact with news and allow for future developments.

Current approaches across several U.S. states, the European Union and China involve implementing disclosure requirements or labeling content that has been generated using AI. These approaches include watermarking, content labeling and disclaimers. The parties responsible for enforcing these types of deepfake regulations have included government agencies but concerns persist that the technology is moving more quickly than legislation and oversight. For many countries, current laws that may be affected by deepfake content (e.g., right to privacy, defamation or cybercrime) do not specifically address synthetic media. These gaps make regulating manipulated content difficult. Enforcement is also problematic in cases where the content creator resides outside of a country’s jurisdiction. 

Experts recommend focusing policies on the general harms of technological innovations rather than on the technologies themselves as it is likely impossible to detect and/or ban all manipulated synthetic media. There are also concerns that deepfake regulation will curtail freedom of speech and expression. Future legislation should consider how to best craft regulations to avoid the costs of such regulations outweighing the benefits. As such, regulations ought to consider the balance between the need for freedom of expression (and an open internet) while also protecting against the harms of mis- and disinformation.

Notable Legislation

Resources & Events

Issue primers have been reviewed at multiple stages by more than 20 global research and industry expert partners, including CNTI advisory committee members, representing five regions. We invite you to send us research, legislation and other resources. Read more about CNTI’s issue primer and other research quality standards.

There are several pieces of legislation, such as the United Kingdom’s Online Safety Act, that are important to consider in broader policy discussions about online safety and algorithmic regulation (and are discussed in other CNTI reports) but they go beyond the scope of this study. This analysis examines the language within 32 “fake news” policies proposed or enacted from 2020 to 2023 in 31 countries, 11 of which have elections scheduled in 2024.

Overall, the study reveals that the language in the 32 pieces of legislation does little to protect fact-based news and in many cases creates significant opportunity for government control of the press. The lack of safeguards in this legislation risks curbing press and journalistic freedoms heading into a major election year. Among the key findings:

• “Fake” or “false” news is explicitly defined in less than a quarter (7/32) of this legislation. Omission of these definitions leaves them open to interpretation by whomever has oversight authority which, in these cases, is often the government itself. 
• In what is very much a double-edged sword, two pieces of legislation examined here explicitly define journalism or what may be considered “real” news, one defines journalists and four define news organizations. While definitions can help protect press freedom, they can also be used as legal grounds to protect media that props up the government and ban media that does not — especially if the court’s application is also dictated by the government. 
• 14 of the 32 policies clearly designate the government with the authority to decide what is or is not “fake news.” In some cases it is the central government itself and in others it is an entity within the government whose independence from the central government is often unclear. The remaining 18 policies provide either vague or no language about who has that control, ceding it to the government by default. Putting this power in the hands of the government — whether explicitly or by default — introduces greater risk of governmental press and message control.
• Although press control issues are more prevalent in the countries with autocratic rather than democratic regimes, definitional issues and a lack of clarity are found in legislation from both regime types. Of the 31 countries studied, 19 are autocratic and 11 are democracies as identified by the research organization V-Dem.¹
• Criminal penalties for the publication of “fake news” vary dramatically, from fines to suspension of publications to imprisonment. Among the 27 policies with clearly noted penalties, three-quarters (20 policies) include imprisonment, ranging from less than one month in Lesotho to up to 20 years in Zimbabwe.

These findings warrant concern. Vague or missing definitions can create, intentionally or not, opportunities for governments to censor opposing voices and restrict press freedom and freedom of expression. Any legislation, even if developed while under leadership that values an independent press, must account for the possibility of future regime change or legal interpretation.

Putting hard lines around false information is certainly not easy. The challenge is exacerbated when trying to discern intentional versus unintentional efforts to mislead. Legislation can be an important part of creating a digital news environment that safeguards both an independent press and the public’s access to fact-based news, but those aiming to develop policy should be aware of these challenges. CNTI offers five key questions that anyone seeking to construct policy to guard against false information in a way that safeguards an independent press and public access to a plurality of fact-based news should consider. Detailed in the “Important Policy Considerations” section of this report, they include: 1) whether policy or other non-governmental methods are the best approach for the current situation; 2) if specific independent oversight is laid out; 3) whether there are clear adjudication processes; 4) who the subject — or target — of the policy is; and 5) what future or global implications might emerge?

This report is one of many CNTI efforts to help address the challenges of today’s digital news environment in ways that safeguard an independent, competitive news media and the public’s access to a plurality of fact-based news. It is also a part of CNTI’s work in the specific area of defining journalism in our digital, global society. Any legislation related to journalism and news needs to thoroughly address the range of implications that can ensue.

Approaches to “Fake News” Policy

A substantial portion of the legislation we examine was developed in response to the COVID-19 pandemic, often targeting the spread of false information about the pandemic as well as information that contradicts government and public health officials. About two-fifths of the policies examined here (13/32) focus specifically on false information about COVID-19 and disputing government public health officials. For example:

• Botswana’s Emergency Powers (COVID-19) Regulations and South Africa’s amendment to its Disaster Management Act use the exact same wording, penalizing “any statement, through any medium, including social media, with the intention to deceive any other person about COVID-19; COVID-19 infection status of any person; or any measure taken by the Government to address COVID-19.”
• The Philippines’ Bayanihan To Heal As One Act penalizes “individuals or groups creating, perpetuating, or spreading false information regarding the COVID-19 crisis on social media and other platforms.”

While a narrow scope might be less susceptible to abuse, these policies nonetheless risk chilling effects on free expression or political criticism. Indeed, narrow “health” proscriptions have been used in Zimbabwe, for example, to persecute journalists questioning COVID-19 policies and exposing corruption in COVID-linked procurement practices. These topic-specific policies also suggest the potential for passing similar legislation in other topic areas deemed risky by the government.

Other recent legislation, meanwhile, include much broader — and more easily exploited — phrases about information that criticizes or harms the country’s military or economy, or sows discord:

• Greece’s legislation on false information includes “anyone who publicly or via the Internet disseminates or disseminates in any way false news that is capable of causing concern or fear among citizens or of shaking public confidence in the national economy, the country’s defense capability or public health.” 
• Hungary’s legislation denotes anyone who “states or disseminates any untrue fact or any misrepresented true fact with regard to the public danger that is capable of causing disturbance or unrest in a larger group of persons at the site of public danger” is guilty of a crime.
• Myanmar’s draft legislation defines the creation of misinformation and disinformation as “causing public panic, loss of trust, or social division on cyberspace.”

Defining “Fake News”

While many countries’ legislation are similar in how they define the type of information that qualifies as false or malicious, the level of specificity varies dramatically across policies, with only occasional shared phrasing (see Table 1). A lack of clarity and use of vague language, when definitional language exists at all, carries risks for journalists and the public.

“Fake” or “false” news (sometimes referred to as disinformation, misinformation or other terminology in legislation) is explicitly defined in less than a quarter of this legislation (7/32). As is the case in disinformation studies more broadly, the question of intent often complicates definitions of “fake news,” with four of the seven seeking to separate accidental or unintentional from intentional spreading of false news (see Table 1).² In some contexts, such as Nigeria, legislation explicitly states that it is illegal to produce “knowingly false” information, though how this definition is adjudicated is unclear.

A natural follow-up question to how false news is defined is if and how journalism — or what might be considered “real” news — is defined. Less than a quarter of these policies define any terms related to journalism. Four of the 32 policies explicitly define news organizations, while journalism (as in news content, not “fake news”) is defined in two of the 32 policies and journalists in only one.

Even when explicit definitions of journalism are present, they are often vague and thereby open to a wide array of interpretations. For example, in Togo’s 2020 Press and Communication Code, journalism is broadly defined as “original content” about current events of “general interest.” It is unclear in this legislation what qualifies as “original” or what counts as “general interest.”

References to or definitions of “journalism” or “journalists” within policy are just as — if not more — critical to fully consider than those of “fake news.” While they may be intended to protect independent, pluralistic journalism, these definitions can also be used to sanctify government control of the press.

Authority to Decide What’s “Fake”

Another element to consider is the authority responsible for, and the process of, determining what or who constitutes “fake news.” This authority is clearly designated in less than half of the legislation (14/32), often making legislation susceptible to abuses aimed at curbing press freedom (see Table 1).

Among the remaining 18 policies with unclear or no noted oversight authority, 11 of them include regulations for “fake news” or disinformation that are wrapped into broader legislation related to COVID-19. The entity most often noted as having any kind of authority within these broader COVID-19 policies is the country’s health minister, but it is unclear exactly what authority that entails when it comes to adjudicating “fake news” specifically.

Among the 14 policies with clearly noted oversight authorities, the arbiter of such decisions is often designated as the head or body of a government commission, ranging from ministers of communications, information or technology to electronic transactions control boards and communications authorities (see Table 1). Others, in practice, have unclear or limited independence from the state. For instance, Togo’s Haute Autorité de l’Audiovisuel et de la Communication (HAAC) has the authority to sanction media actors and grant press accreditation. While its structure is technically independent from the government, it is not particularly transparent and in the past has been found to censor news organizations at the state’s request.

Penalties

The criminalization of “fake news” has a long history in many countries that have experienced penal codes imposed under colonization. Similar types of penalties are embedded in many of these modern policies.

About four out of five of these policies (27/32) explicitly note criminal penalties for creating and publishing “fake news” or disinformation — most commonly fines and/or imprisonment, though some legislation includes the temporary suspension of news publications (e.g., Togo) or compulsory community service (e.g., Uzbekistan).

Among the 27 policies with clearly noted penalties, three-quarters (20 policies) include imprisonment, ranging from less than one month in Lesotho to up to 20 years in Zimbabwe.

Together, these policies reveal the serious consequences journalists face especially when legislation does not include clear oversight authorities or processes and is then abused by governments. Further, the risk of imprisonment or heavy fines may create broader impacts, discouraging independent journalists through  potential consequences such as their work being labeled as “fake news.”

Important Policy Considerations

While news and news organizations are not necessarily a central consideration of “fake news” or disinformation policies, the ripple effects of these policies impact independent journalism and press freedom. Codified policies focused on news, disinformation and/or journalism — as well any language used to define these and related terms — carry tremendous power, creating opportunities for governments or other powerful actors to intentionally or unintentionally threaten the independence, diversity and freedom of the press as well as broader elements of free expression.

As Table 1 shows:

• A majority of the countries examined in this study (19/31) have some form of autocratic governance (versus 12 democracies).
• Among the 30 countries in this study that also appear in the Reporters Without Borders Press Freedom Index, more than half (17/30) rank in the bottom half of the index. 

Together, these findings indicate high risks of undue political influence compounded by vague language and a lack of clear definitions and oversight processes. They also suggest that a majority of these policies may not have been intended to promote democratic values such as an independent press or free expression. Those aiming to construct policy that promotes an informed public and independent press should be aware of these conditions.

It is critical to take a careful and deliberate approach to codification of what “fake news” (real news) is and is not. This is not an easy task, as CNTI discusses in some of our issue primers, and has become even more complicated in the digital era. This study reveals five important questions to consider in  policy development in this area:

• Are legislative policy or other non-governmental methods the best approach to address disinformation? The latter could include processes such as human rights or information risk assessments, content moderation or data sharing on digital platforms, news and media literacy initiatives or other related efforts. Each option, including legislative policy, has potential benefits and harms which are important to think through.
• Are specific, independent oversight of these definitions included in policy? This can be designated via self-regulatory bodies or through independent government agencies designed to protect against undue political influence.
• Are there clear adjudication processes for these definitions? Can journalists, news entities or civil society formally challenge definitional decisions by oversight bodies and if so, how?
• Who is the subject of the policy? Who would be liable? Solely individuals? Publishers? Platforms? Some combination based on the circumstances? While complex, it is important to fully consider who the law would implicate and why.
• What potential future and global implications might emerge? These decisions have global consequences, as policies in one country inherently impact those in others. And, as new technologies for disinformation and digital manipulation emerge, new tactics for addressing them may be necessary. 

The Issue

The internet is enmeshed in virtually every aspect of modern life. Most often, the public’s online experiences flow through digital platforms, which serve as intermediaries between the content being delivered and the individuals consuming it. That provides convenience but also raises serious concerns. A central one is how much control digital platforms have, or should have, over what people see, especially when it comes to news and information, and how transparent digital platforms’ decision-making processes, typically conducted through the use of algorithms, needs to be.

Each digital platform, and its algorithmic infrastructure, operates under its own rules and conventions to determine which content is and is not allowed as well as what content is prioritized on the platform. Algorithmic choices have far-reaching implications ranging from determining user experiences by selecting one news story over another to facilitating hate speech, disinformation and violence. 

Policymakers, journalists, social media researchers, civil society organizations, private corporations and members of the public are raising awareness of and questions about what are often-opaque digital platform practices related to content selection and moderation processes. The global news industry in particular – which relies to a significant degree on digital intermediaries to reach audiences – has articulated the need for a better understanding of how algorithms rank, boost, restrict and recommend content (including, but not limited to, news) and target consumers. Without a better understanding of how these algorithms are currently built, it is difficult for publishers to work within their constraints to reach audiences or to play any role in determining how algorithms should be built to promote an informed public.

Thus, in addition to ensuring that algorithms are designed to identify fact-based, independent journalism (a topic CNTI addresses in a separate issue primer), it is critical to enhance transparency in and accountability for how digital platform algorithms function.

The key challenge is that regulating or mandating transparency is more complex than it appears, with many elements that need to be thought through. What forms of transparency are necessary and most likely to be effective for protecting and promoting access to high-quality, diverse and independent news? What does transparency mean to policymakers, journalists, researchers and the public? How can we enable algorithmic transparency in ways that protect against political and commercial manipulation or abuse and ensure user privacy? What are its limitations? What systems or entities should be beholden to these levels of transparency? And who should be empowered to set these rules? 

Addressing the questions introduced in this primer is a critical first step toward developing approaches to algorithmic accountability that support an independent, competitive press and an informed society. Policymakers and other stakeholders can benefit from a more nuanced understanding of what transparency can, and cannot, accomplish as well as what new risks could be introduced. Policymakers must be forward-thinking about how and where algorithms may be used in the future, what data could be exposed and who could ultimately gain access to it as a result of increased transparency.

What Makes It Complex

One hurdle in creating valuable algorithmic transparency is being sure that those on the receiving end are armed with the necessary knowledge to assess these processes. A good deal of tension has arisen over how technology companies have responded to requests for independent researcher, journalist and public access to algorithmic processes or their digital trace data. Transparency policies must take into account each stakeholder’s differing needs, aims and interests and consider what forms of transparency are necessary and appropriate for each. Because the algorithmic processes are complicated and often rely on technological jargon, there is the risk that even with increased transparency, a lack of understanding of these processes will endure for some stakeholders.

“Open-source” transparency is not enough to achieve an understanding of algorithmic processes or solve the problems of platform accountability. This is true in two different manners. First, some tech leaders have suggested “open-source” algorithms as an answer to transparency and accountability issues, but algorithmic processes require a level of technical knowledge most stakeholders do not possess in order to understand their complexities. “Open-source” transparency requires individuals to have the time and capacity to learn the complexities of data governance and content moderation. Context is critical in learning how algorithms work; without a clear understanding of the internal policies dictating algorithmic behaviors or access to their underlying training data, raw information is not interpretable.

Determining the parameters of transparency is challenging to do in a single policy. There is broad agreement that transparency is a critical component of digital platform governance and content moderation, and that “more transparency is better,” but, it is not always clear what, and where, transparency is being sought. Collaboration between experts is critical to addressing the multifaceted impacts policy on transparency may have. Further, the aim of transparency differs in market-based versus state-led models, with the former aiming to empower users and the latter aiming to empower regulators. Policymakers, researchers, the news media and civil society must work together to weigh stakeholders’ needs against platforms’ capabilities and consider what regulatory structures will serve to enforce this legislation. Finally, as we address in a separate issue primer, addressing transparency on its own does not address disagreements over how to ensure people see fact-based, independent news content on digital platforms, nor does it address the choices individuals themselves make about what they click on or what kind of information they seek out.

Digital platform regulation needs to protect against political and commercial manipulation as a consequence of algorithmic transparency. As discussed in our issue primer on addressing disinformation, algorithmic transparency (and researcher access to digital trace data) can both help experts understand the spread of online disinformation and how to counter it and, at the same time, introduce new risks for manipulation and abuse. Governments in many parts of the world increasingly pressure platforms to make content decisions that benefit their own political interests. There is also the risk that commercial actors or institutions manipulate algorithms to exploit virality, amplifying large quantities of clickbait and problematic content over fact-based, independent information – particularly by using new technologies like generative AI. Both introduce threats to an independent press and an informed public. Policy must safeguard against these risks and insulate regulation from political and commercial manipulation, both by ensuring oversight remains independent from governments and by separating regulatory bodies from direct involvement in content moderation decisions.

Platforms’ commercial structures are real factors to be considered, both in the way they can negatively drive internal decisions and in the value they bring to our digital environment. Technology companies are private business entities that operate in a competitive marketplace with a certain level of value placed on their proprietary information and intellectual property. The current concealed nature of algorithmic selection and ranking, however, risks commercial incentives negatively affecting the work of journalists and the content the public receives. Crafting the best transparency policies will require a balance between these two elements, including consideration of potential unintended consequences for innovation or competition, such as deterring startups or smaller players from participation due to costs associated with adhering to regulation.

Debates around transparency introduce important opportunities and risks for the relationships between governments and platforms. This is particularly important when it comes to who oversees or regulates algorithmic transparency and what can be asked of platforms when it comes to content moderation. Legislation that does not address these critical questions can threaten an independent press and freedom of expression (for instance, via “jawboning” practices or other legal demands of platforms). This is a consistent complexity across many of CNTI’s issue areas and speaks to the importance of considering the balance between legislative and organizational policy (including government-enforced organizational policy) in addressing these challenges.

Policy must balance transparency and accountability against user privacy concerns when applicable. Digital platforms’ efforts to improve transparency, when they exist at all, are at times negotiated within the contexts of user privacy concerns and data security concerns. While this does not affect all approaches to algorithmic transparency, questions about what user data digital platforms would be obligated to share and what personal information would be contained in that data will inevitably lead to trade-offs between digital platform transparency and user privacy.

State of Research

Over the past decade, questions surrounding digital platform governance have moved to the forefront of political, legal and academic debates. Recently, a breadth of interdisciplinary research has focused on social media platforms and messaging apps central to global journalistic work as well as on the broader range of software companies and sharing or gig economy apps central to contemporary online commerce. 

Much of this research fits into two categories: governance of social media and governance by social media. Research focused specifically on digital platform algorithmic transparency represents one small segment of work in this field. 

To date, this research has been far more theoretical than empirical, largely due to a lack of data (as we note in greater detail below) and the slow-moving nature of legislative policy, but it still sheds light on the varying forms of transparency different stakeholders expect. For instance, public transparency via disclosures looks different from, and accomplishes different aims than, research transparency via data access. 

Research findings have also noted the limitations of algorithmic transparency alone as a means of establishing digital platform accountability as well as the challenges various legislative efforts face in addressing algorithmic transparency, ranging from potential free expression violations to individual privacy infringements. These limitations have led some experts to call for forms of digital platform accountability beyond algorithmic transparency. 

Collectively, this work also reveals how little has changed, both via research and policy, in almost a decade of calls for algorithmic accountability.

Looking forward, there is much that we do not yet know about digital platforms’ algorithmic infrastructures. Research in this area is critical to inform current debates about internet governance. Experts have called for independent researchers and civil society to be granted more access to the inner workings of digital platform technology companies, the algorithms they develop and the trace data they collect – particularly as digital platforms steadily move away from more open models of free data sharing through their APIs. Of course, what access to give to whom is itself part of the debate about transparency. Still, access, when it has the appropriate safeguards, can help to better inform policy-making and contribute to democratic processes as a form of checks and balances on government and corporate power.

State of Legislation

Content moderation is the element of transparency that has, to date, received the most policy attention. A wide range of policy initiatives around the world have begun, albeit slowly, to regulate content moderation processes and/or government takedown requests on digital platforms, which introduce opportunities and risks. We address some of these in a separate issue primer. 

Legislative policy around addressing algorithmic transparency and accountability more broadly is still nascent. While some technology companies have taken steps internally to improve organizational transparency, governments are beginning to consider policy initiatives that require specific forms of platform accountability, with a focus on (1) supporting research access to data, (2) protecting user data privacy and (3) disclosing certain content moderation practices. 

Some experts have cautioned against the risk that poorly designed transparency laws could become a mechanism for state intervention in and control over platforms’ editorial policies. In contexts like the U.S. with strong constitutional protections of free expression, it is unclear whether the government can mandate transparency. 

The establishment of international ethical frameworks for platform accountability is complicated by the fact that most institutional structures for oversight, such as review boards or ethics committees, vary greatly by country. Nonetheless, policymakers, particularly in Europe and the U.S., have called for international cooperation when it comes to facilitating transparency and access to cross-platform research. Experts have also called for multi-stakeholder and (supra-)national legislative efforts to govern digital spaces.

Even if not intentional, policymakers need to consider that a legislative framework in any one country has the potential to influence regulation globally. For example, a rule-of-law-based approach, while effective in stable democracies, could later serve as a blueprint for suppression elsewhere.

Notable Legislation

Resources & Events

Issue primers have been reviewed at multiple stages by more than 20 global research and industry expert partners, including CNTI advisory committee members, representing five regions. We invite you to send us research, legislation and other resources. Read more about CNTI’s issue primer and other research quality standards.

The Issue

The display of news content on digital platforms deviates from the traditional news model by decoupling news production from news distribution. In addition to news organizations’ own digital distribution, technology companies’ algorithms distribute news by selecting, filtering, ranking and bundling news for consumers. News organizations increasingly depend on digital intermediaries to reach audiences through these means, and the public, in turn, relies on digital intermediaries to access news. Social media platforms, search engines, news aggregators and video-sharing services are becoming the dominant means for news consumption across the world. While these pathways increase news access and reach, they also introduce two challenges: the algorithmic selection process is usually opaque to news publishers and the public (who both rely on it), and the algorithmic selection results have the potential to expose the public to lower-quality news and information.

In addition to addressing transparency in how digital platform algorithms function (a topic CNTI addresses in a separate issue primer), there is a need to understand how to ensure fact-based, independent journalism rises to the top on digital platforms. Both publishers and digital platforms face challenges over how to determine the newsworthiness of editorial content. In particular, how can algorithms find, select and prioritize accurate, evidence-based content? This is important in enabling:

• An informed public. News reporting remains critical to supporting an informed public. With the digital environment open to both well- and ill-intended actors, algorithmic selections risk delivering lower-quality or erroneous content to the public. In addition, it is possible that amid declining revenues, some publishers may attempt to prioritize and elevate content selected and amplified by platform algorithms over coverage driven by public interest.
• Journalism’s sustainability. News publishers increasingly rely on digital platforms to reach a broader audience, to drive more traffic to their own websites or apps and to increase subscriptions and donations. Less than a quarter (22%) of news audiences say they prefer to start their news journeys with a news website or app, down 10 percentage points since 2018. Instead, social media platforms play an increasingly critical role in news consumption. Amid efforts to seek more revenue from algorithmically driven news consumption (discussed in a separate CNTI issue primer), the opportunities digital platforms provide for exposure and reach to fact-based, independent content remain critical for many publishers. There is also the real possibility of platforms shifting away from news altogether which, at least in our current structures, would almost certainly harm publishers’ ability to reach audiences and inhibit the public’s ability to access legitimate news sources and information. 
• Digital platforms’ relevance, trust and revenue. Although news content typically makes up only a small portion (according to Facebook, around 3-4%) of digital platform content, platforms that provide independent and fact-based news content reap the benefits of increased relevance, credibility and some revenue when they display ads next to links or snippets and collect user data for targeted advertising. News generally improves the quality and range of content on digital platforms, whose aim is to entice users to spend as much time as possible in their “walled gardens.” 

Behind the challenge of ensuring that algorithms identify and promote fact-based, independent journalism, is the influence digital platforms possess through subtly guiding how society creates and discovers content. Digital platforms have evolved beyond the role of distribution channels to exert direct and indirect editorial influence (though some have begun to shift away from news). For example, in the past decade, digital platforms have incentivized particular types of content (e.g., live video, short video) and set design standards (e.g., subscription policy, anti-cloaking). As a result, editorial decision-making in a competitive, data-driven news market is increasingly based on third-party choices, resulting in newsrooms optimizing online editorial content for clicks, shareability and engagement. 

These dynamics signal the growing importance of ensuring that platform algorithms are structured in a way to identify and promote fact-based, independent journalism. There is a clear need for both legal and organizational policy when it comes to algorithmic transparency. There is also a critical need for cross-industry conversation and collaboration. There is evidence that collaboration leads to increased visibility of fact-based journalism by, for example, changing algorithms (and including human validation) to prioritize and elevate original reporting over standard ranking considerations. Is there a role for news publishers in shaping these algorithmic processes? And, what do publishers need from digital platforms to be able to serve the public and make effective and informed decisions about digital news content?

What Makes It Complex

Policy in this space will need to grapple with what quality means and how it gets assessed. What content algorithms should promote, as well as who or what defines news “quality,” can be difficult to determine or to articulate in policy. And it often means different things to academics, to news publishers and to the public.Historically, the term “quality” was used by broadsheet newspapers and public broadcasters to distinguish their coverage from the more quantity-driven, sometimes sensationalist tabloid headlines as well as from commercial channels – though whether this distinction was ever true is questionable. It can be difficult, if not outright impossible, to label many news publishers entirely along a binary measure of “quality.” (Further, research suggests that such labels don’t measurably improve people’s news diets or reduce misperceptions.) In some contexts, the question of who determines the quality of news content can open the door to governmental overreach or abuse. Governments’ involvement in decisions around what qualifies as “quality” news content can lead to them serving as the arbiters of what does or does not appear online, potentially threatening an independent press and free expression.

Even if algorithms prioritize fact-based and independent journalism, people may still choose to consume content that is not fact-based or independent. For instance, tabloid publishers have often argued their popularity guarantees that their content is what the public wants. The disconnect between what information publishers and the public deem relevant or valuable is a challenge for legal and organizational policy attempting to address algorithmic content recommendations. For instance, should the public be compelled to see content they don’t want to see if governments or others determine it is good for them? Efforts must strike a balance between prioritizing fact-based content and protecting the public’s right to make their own decisions about how, or to what extent, they are informed. This disconnect is explored in more depth in a separate CNTI issue primer on news relevance.

Even if we can agree on the standards of quality journalism, the challenge of how to choose among sources remains. For instance, many news stories repeat, repackage, aggregate or comment on previously published content. How can policy or other approaches disincentive these practices in favor of original reporting? What role might platforms perform in this? Further, in today’s attention economy, as algorithms sift through and prioritize content from an ever-growing number of news publishers and freelancers, smaller brands are often at a disadvantage amid the prioritization of publishers’ reach and quantity. Further, because of their assumed public and economic value, legacy and upmarket publishers often have more access to digital platform and search engine representatives and may receive more guidance on what types of content will perform better in algorithms. Meanwhile, smaller and often local news organizations may not be able to afford the same level of digital expertise needed to best navigate news algorithms. A large proportion of digital subscriptions already go to just a few big national brands, reinforcing a “winner takes most” dynamic in the online media industry.

Much like their audiences, news publishers generally have a limited understanding of how algorithms target news consumers and rank, boost, restrict and recommend news content. To journalists, the metrics used in these algorithms range from opaque to obscure, despite some efforts to improve transparency. When proprietors strategically make regular and unannounced changes to their algorithms, ensuing decreases in traffic often lead newsroom, SEO and social media teams to feel at the mercy of major technology companies. These knowledge gaps make it difficult for news publishers to participate in decisions about how platform algorithms should be built to promote an informed public by recognizing and encouraging fact-based, independent journalism.

Platform algorithms are entangled with broader human choices and user preferences, so they are not necessarily designed to prioritize news content. Platform algorithms are generally built to respond to people’s broader interests, which may or may not include news and current affairs. Individuals see customized search results and feed pushes, and it is unclear whether the public desires more uniform results and pushes.

State of Research

The impact of algorithms on news media has been recognized as a critical area of research. The central academic debates focus on the risks and opportunities algorithmically ranked news presents to individuals and society, digital platform accountability and the shift from mass media distribution to personalized recommendation systems. Scholars across a range of fields have increasingly analyzed the critical relationship between those who manage algorithms and those who produce news. 

Researchers have found evidence of publishers adjusting – even lowering – editorial standards to boost their position in ranking and recommendation systems. Researchers have also expressed concerns about accessing, understanding and evaluating the algorithms that govern a range of news-related processes. These concerns involve the opaque, “black box” nature of these algorithms as well as the increasing suppression of researchers’ access to digital trace data (which CNTI addresses in detail in a separate issue primer).

While it is clear that platform algorithms subtly guide trends and drive how people discover content, there is little evidence of exactly how they work or what they reward or amplify. Thus, it is important to differentiate between what the data firmly reveal about algorithms and news and what there is not yet evidence to support:

• In spite of public attention to fears of algorithmic “filter bubbles,” studies in the UK and several other countries consistently find that algorithmic selection by digital platforms generally leads to slightly more diverse news consumption (though self-selection may hinder this, particularly among political partisans). 
• Algorithmic rankings can exert influence over how people engage with and evaluate news.
• Some work suggests that emotional language and out-group animosity are more likely to go “viral” or be shared on social media – trends that rely on user behavior alongside platform and algorithmic design. 
• Historically, social media platforms like Facebook overwhelmingly weighted comments as more important than other types of interactions, leading posts that explicitly or implicitly encouraged commenting (e.g. divisive content) to gain more popularity. But transparency around algorithms has limitations because the code is not static: Platforms’ formulas for engagement are frequently changing.
• Research demonstrates that platforms are frequently changing algorithms and investing in both manual and algorithmic moderation processes to minimize the spread of disinformation. At the same time, problematic or extremist content is more likely to be amplified by certain digital platforms’ recommendation systems than by others.
• News consumers are generally skeptical of algorithmic selection of online news but are equally wary of news selection by editors and journalists. Similarly, people often perceive news produced by algorithms as equally or more credible than (or simply not discernible from) news selected by human journalists. 

There is opportunity for future work to serve as an intermediary between digital platforms and news organizations, pushing for more transparency to fully understand the complexity of algorithmic mechanisms and where the balance may lie between demystifying processes to promote fact-based, independent news and information while protecting against political or commercial manipulation. On the policy side, the effects of recent digital platform legislation on newsroom practices and reach are unclear. Establishing best practices, both in digital platforms and editorial environments, will depend on the findings of research in this area.

State of Legislation

Legislative approaches to address platform algorithms in regard to news content and distribution range from direct government intervention to self-regulation. While laws and proposals vary by country, they face a common challenge: the “pacing problem” – technology moving faster than the law.

We discuss two key groups of approaches addressing the impacts of algorithm-driven services with policy in separate issue primers. The first group appears in countries where the emphasis is to protect citizens against safety risks and to increase algorithmic transparency and accountability. The second group appears in countries like Australia, Canada, Brazil and the U.S. where the aim is to require large platform companies to negotiate commercial deals with news publishers for their content. Within both sets of approaches, it is unclear how to assess what “quality” news is. 

Legislation in these areas can greatly impact – and even supersede – what news publishers and technology companies can do to promote fact-based, independent content. In some cases, media bargaining structures, which often primarily benefit legacy and larger publishers, would have determined these actions. Thus, legislation may end up being the arbiter of what content appears online. Empirical evidence is needed to better understand how regulation impacts the algorithmic information landscape in practice, with intended and unintended consequences for fact-based editorial content and news consumption.

Notable Legislation

Resources & Events

Issue primers have been reviewed at multiple stages by more than 20 global research and industry expert partners, including CNTI advisory committee members, representing five regions. We invite you to send us research, legislation and other resources. Read more about CNTI’s issue primer and other research quality standards.