Introduction: Information landscape and humanitarian contexts

The spread of false information – whether overt or covert – poses a serious threat to the humanitarian sector, undermining its work and endangering the populations it aims to serve. While not new, digitalization has transformed the scale, speed and complexity of harmful information. This shift is unfolding amid growing distrust in institutions, making its effects even more corrosive. Today's information environment is crowded and layered, with multiple forms of harmful content coexisting and reinforcing one another. Harmful information rarely exists in isolation; it amplifies other risks such as geopolitical tensions and environmental crises. The World Economic Forum's Global Risks Perception Survey 2024–2025¹ warns that vulnerabilities linked to online activity are deepening alongside widening societal and political divisions, eroding public trust in information and institutions.

Humanitarian contexts are inherently complex, particularly during disasters, armed conflict and other emergencies. In such situations, the state plays a central role in shaping the information environment – serving as the primary source of official communication, issuing alerts, coordinating messaging and engaging with both domestic and international responders. Timely, accurate and trusted information is essential: it saves lives, fosters trust and supports public order. However, malicious actors often seek to control or distort these narratives. Countering such efforts requires not only credible channels of communication but also organizational and societal resilience.

An important actor navigating this complexity at the national level is the National Red Cross and Red Crescent Society. National Societies hold a unique position as auxiliaries to public authorities in the humanitarian field – a status that allows them to support and engage with the state while maintaining independence and adherence to humanitarian principles.

The status of National Red Cross and Red Crescent Societies is not arbitrary; it is formally recognized by states that are signatories to the Geneva Conventions. These states have committed – through both the UN and the International Conferences of the Red Cross and Red Crescent – to uphold and respect the Movement's fundamental principles. These principles – humanity, impartiality, neutrality, independence, voluntary service, unity and universality – are essential to ensuring that National Societies can operate effectively and free from political interference. In practice, this means that National Societies must be able to deliver humanitarian assistance independently, impartially and neutrally, even in complex or politically sensitive environments. Respect for these principles is not only a matter of good practice; it is a legal and moral obligation for states that have endorsed the Geneva Conventions and the commitments made in international humanitarian forums.

Traditionally, this auxiliary role has encompassed disaster preparedness and response, health services, support to vulnerable populations and the promotion of humanitarian principles and international humanitarian law. In practice, it also demands trusted public awareness, communication of information and meaningful community engagement – essential to effective humanitarian action. National Societies contribute in multiple ways: communicating early warnings and public health messages; coordinating with authorities to ensure consistent, timely and localized information; supporting risk communication and community feedback mechanisms; and increasingly working to counter harmful information and rumours that may undermine humanitarian responses.

Given the complexity of today's information environment, there is a need for greater recognition of – and support for – the independence of National Societies in fulfilling these roles. The 2024 Council of Delegates called on each National Society to strengthen dialogue with public authorities to reinforce their independent action and decision-making, including by anchoring their auxiliary role in domestic law in line with Movement standards and past International Conference resolutions.²

5.1 Defining harmful information: A strategic and contextual challenge

Determining what constitutes harmful information is essential for any entity seeking to develop a strategic response aligned with risk management. The impact of harmful information depends on several variables including its scope (scale and severity), duration, magnitude of the incident and the resilience of the affected individual(s), organization(s) and broader context. Harm may result directly from the incident or indirectly, with distinctions often based on the degree of certainty that the information caused the outcome. A single incident can generate multiple forms of harm, classified as direct (primary) or indirect (secondary or tertiary). These may include physical, psychological, societal and deprivational harms. The effects of harmful information are often disruptive and multidimensional, manifesting across several categories simultaneously.

Governments interpret harmful information in diverse ways, shaped by their legal frameworks, political priorities and societal values. Some describe it as deliberately false or misleading content intended to deceive, manipulate or cause harm – whether to individuals, public institutions or national security. Others place greater emphasis on intent, such as its use to influence elections, incite violence or undermine public health measures.

Common definitions distinguish between:

• disinformation: deliberate falsehoods intended to cause harm

• misinformation: false or inaccurate information shared without harmful intent

• malinformation: genuine information used with the intent to cause harm.

In these framings, intent is the key differentiator.

One of the most contested areas in defining harmful information is the boundary between legitimate political dissent and incitement. In some contexts, this line is blurred, raising serious concerns about freedom of expression. Content may be labelled as dissenting, destabilizing or inciting violence, particularly during times of crisis or internal tension. The challenge is ensuring that measures designed to counter harmful information are not used as a pretext for silencing dissent. Clear safeguards are needed to distinguish between speech that challenges authority and speech that genuinely threatens public order, safety and human dignity.

You know, … I don't think the law has caught up with social media and being able to kind of hold people to account for the information that they've been putting up there. And I think, … you know, we've got a right to kind of free speech. So … there's a conflict there, isn't there? To a certain extent, you know, he can say anything he likes, but it could be false… And then there's the information that people have a right to know, what the truth is. So I think there's a difficult balance to get between the laws and free speech."

5.2 The risks of information control in emergencies

While states have a legitimate responsibility to protect territorial integrity and public order, overly broad or opaque security measures can come at the expense of individual rights and community safety. In some contexts, authorities have withheld or delayed the release of critical information, shared incomplete or misleading narratives or used public messaging to shape perceptions during emergencies. In situations of armed conflict, information often becomes a strategic instrument – used to influence, mobilize or obscure.

Measures such as restricting, blocking, filtering, censoring or regulating content are sometimes introduced under the pretext of national security, public order or preservation of culture. Yet, when vaguely defined or applied without transparency, such measures can undermine access to essential information, restrict freedom of expression and compromise the integrity of the information space. Those with disproportionate control over media and platforms can use harmful information to discredit civil society and humanitarian organizations, associating them with malign or unlawful actors or blaming them for crises to justify repressive policies. This manipulation fuels discrimination, human rights abuses and social tensions.

At the core lies a critical tension: addressing harmful information effectively while safeguarding fundamental rights. Vague or overly broad definitions of harmful information risk misuse – suppressing journalism, silencing dissent or curtailing civil society. In some contexts, legislation intended to combat disinformation has been used to detain journalists, restrict reporting or target individuals for expressing political opinions. Such measures can erode public trust and deepen fear and polarization. Effective responses to harmful information must therefore be grounded in legality, necessity and proportionality.

A body of UN human rights instruments, resolutions and guidance underscores the importance of protecting human rights online, including freedom of expression and access to information. These instruments provide essential guidance in navigating the tensions between addressing harmful information and safeguarding fundamental rights. For example, Article 19 of the Universal Declaration of Human Rights (1948) states that "Everyone has the right to freedom of opinion and expression; this right includes freedom to hold opinions without interference and to seek, receive and impart information and ideas through any media and regardless of frontiers."³

One notable multilateral initiative is the Freedom Online Coalition, a group of 42 states committed to advancing internet freedom. The coalition has repeatedly urged governments to refrain from sponsoring disinformation, including campaigns that undermine humanitarian principles or incite violence. It also emphasizes the importance of maintaining a global, free, open, secure and interoperable internet and other digital communications services during times of armed conflict. The coalition has warned that internet shutdowns hinder access to life-saving information for crisis-affected populations, disrupt protection mechanisms and vital services and obstruct the delivery of humanitarian assistance.⁴ Importantly, this underscores not only the dangers of disinformation but also the risks posed by the absence of information – a critical concern in emergencies where timely communication can mean the difference between life and death.

5.3 Digital ceasefires and harmful information

While traditional ceasefire agreements focus on halting the use of force and physical violence, there is growing recognition that the information space – including the spread of harmful information – can also threaten the sustainability of peace. Harmful information, such as hate speech, incitement to violence and disinformation, have been used to undermine ceasefires, discredit negotiation parties or provoke renewed hostilities. In some recent contexts, mediators and peacebuilders have begun to acknowledge the strategic role of information and have advocated for explicit clauses in ceasefire and peace agreements that explicitly reference the prohibition of hate speech, incitement and disinformation, as well as the importance of access to the internet and accurate information. However, such provisions remain rare, vague or insufficiently monitored.

Incorporating information-related commitments in ceasefire agreements, such as respecting media freedom, refraining from online incitement and establishing joint communication mechanisms, could strengthen trust between parties and help prevent a relapse into violence. This is particularly critical in contexts where digital platforms are actively used to mobilize support, reinforce polarization or spread disinformation.

5.4 Sovereignty in cyberspace

Determining whether an influence operation in cyberspace is a breach of sovereignty is a complex issue under international law. It is closely tied to whether the act infringes on a state's territorial integrity or interferes with inherently governmental functions, such as delivering public services, conducting elections or collecting taxes.

Discussions at the UN Open-ended working group on responsible use of information and communications technologies (ICTs) have examined these issues in the context of cyber incidents.⁶ Key considerations include whether an incident damages or limits the functionality of infrastructure or related equipment; alters or deletes data; interferes with inherently governmental functions; and whether a state has sought to influence, disrupt or delay democratic processes in another state which could include through propaganda, disinformation or covert actions.⁷ Thus, whether such acts amount to a violation of sovereignty depends on their nature and repercussions: Does the incident violate territorial integrity? Does it interfere with or usurp an inherently governmental function?

Researchers such as Pamment emphasize the importance of assessing foreignness in influence operations, specifically, whether they have connections to foreign states, citizens or interests aiming to influence public opinion in another state.⁸

5.5 Media as a pillar of societal resilience

The level of support – or lack of – that states and other stakeholders show for the presence of humanitarian organizations on their territory can significantly influence how they are perceived. When an organization is portrayed as biased, politicized or pursuing its own agenda, its legitimacy is undermined, particularly in contexts where there is limited public knowledge or familiarity with its mandate or work. In such environments, perceptions matter deeply. Humanitarian organizations must be seen as principled, competent and effective. Where public awareness is low and/or where polarization is high, people are more susceptible to harmful information, making them more vulnerable to misconceptions and distrust when forming their opinions.

Researchers Humprecht, Esser and Van Aelst identify five key risk factors that influence a society's vulnerability or resilience to disinformation.⁹ These are: (1) high levels of societal polarization; (2) low trust in news media; (3) highly distributed media landscapes (offering more entry points for disinformation); (4) large media markets, where attention-based revenue models incentivize sensational or false content – what the Global Disinformation Index describes as "the loudest voices getting the most attention"; and (5) high levels of social media use, which are consistently correlated with greater susceptibility to disinformation.

According to the Reuters Institute Digital News Report 2024, audiences are drawn to video and other content on social and video platforms for three main reasons: the perceived authenticity of unfiltered, user-generated content, which appears more trustworthy; a preference for making up one's own mind without editorial framing; and growing mistrust in traditional media.

Many users, especially younger audiences, tend to trust bystander or first-person footage more than traditional news sources, perceiving it as less filtered, less biased and/or less politically manipulated. Concerns about misinformation often centre less on entirely false content and more on seeing opinions and agendas that they may disagree with, perceived bias or superficial, unsubstantiated journalism.¹¹

These dynamics underscore the importance of visibility, transparency and trust-building in humanitarian communication – especially in environments where disinformation can fill gaps left by limited public understanding.

Contributor Insight 5.3

Audiences increasingly rely on social media for news despite concerns about information quality

Social media and video networks have become increasingly central to how people access news in recent years, while other forms of news consumption have dwindled. Data from the Reuters Institute Digital News Report 2025 show that social media overtook television as a source of news in the US for the first time, and in many Latin American, Asian and African countries, social media has been the main source of news for some time.

Generational shifts are driving much of this change. While older adults (55+) continue to rely on traditional media for news, everyone else is now effectively digital-first. Within that online space, younger groups have become less likely to go directly to a news website or app – and more likely to consume news via social media or video networks. Across the 48 countries surveyed, 44% of under-25s and 38% of those aged 25–34 prefer accessing news through platforms, while just a quarter of each group prefer going directly to news websites. Respondents cite convenience and relevance of the news they see through platforms as key reasons – it is often encountered while doing other online activities.

Fig 5.2

Proportion of each age group that say each are their 'main source' of news, 2025

Aggregate data from 48 countries

Fig 5.3

Proportion of under‑25s who prefer to access news via websites or social networks, 2015–2025

Aggregate data from 48 countries

The specific platforms used for news have also changed over time. While Facebook remains the main social network for news, its use has declined considerably since 2016. Meanwhile, video-based platforms such as YouTube, Instagram and TikTok have grown in popularity, resulting in more fragmented attention. Across 12 countries tracked since 2014, six networks now have a weekly reach of 10% or more of the population, compared to only two platforms a decade ago.

Fig 5.4

Proportion of respondents who used each as a source of news in the past week, 2014–2025

Selected countries

Across these crowded digital platforms, news organizations must compete for attention with a range of other voices, many of whom are often more effective at attracting audiences through partisan commentary, engaging story-telling and relatable personalities often perceived as more authentic.

Our data show that traditional news media particularly struggle to retain audience attention on video platforms, where algorithmic recommendations – rather than, e.g., the suggestions of friends – play a dominant role. While audiences still tend to pay more attention to traditional news media and journalists on Facebook, online creators and personalities now outperform established news media on platforms like TikTok and Snapchat.

Even as audiences are embracing platform-based consumption for news, many express concern about the reliability of the information they find there. Facebook (the most used platform for news) and TikTok (the fastest growing platform for news) are seen as the biggest problems when it comes to misinformation, with around half of global respondents perceiving each as a major threat. Although trust in news overall has declined in many countries, findings repeatedly show that traditional news media are still trusted much more than social media or search platforms when it comes to important news.

Across markets, 58% of respondents say they worry about distinguishing true from false news on the internet – a concern that is much higher in Africa and parts of Latin America. When people need to check information that they suspect is false, the most widely cited source is a 'news brand I trust' (38%), ahead of official sources, search engines and fact-checking websites. Even among those who say they would verify information via social media and search engines, it is 'trusted news brands' that people are most likely to be using. In most countries, these trusted sources tend to be the brands (and websites) of news organizations with a reputation for impartial news such as the BBC in the UK, ARD in Germany and NHK in Japan.

Fig 5.5

Top 3 news brands used to check whether something is true

Trusted brands

Concerns about (hyper) personalization, polarization and misinformation and disinformation are likely to intensify with AI, as synthetic content floods the internet and more audiences use AI to access information. Among respondents, 7% already use an AI chatbot for news on a weekly basis – rising to 15% among those under-25. The integration of AI into widely used tools like search engines may drastically increase the uptake of these technologies, even if users are not always aware they are interacting with AI. Across countries, survey respondents believe AI will make news cheaper to produce (+29 net difference) and more up to date (+16), but less transparent (-8), less accurate (-8) and less trustworthy (-18). These concerns could increase the value of trusted news brands, including their websites and apps, which may serve as reliable anchors for audiences seeking information on contentious and important stories.

Fig 5.6

Net difference between proportion of respondents who think generative AI will make news more or less of each

Average score across 31 countries

The outlook for news organizations remains uncertain. Platform- and algorithm-driven access to news is now the dominant behaviour, especially among younger audiences, significantly weakening the traditional so-called 'post and refer' model, as platforms increasingly try to retain users within their own ecosystem. The rise of generative AI search interfaces is expected to further challenge efforts to draw audiences back to news websites. However, journalism continues to play an important role in the information environment, especially in times of crisis. To remain relevant, publishers will need to rethink how they reach audiences and develop more sustainable business models.

Looking ahead, a central challenge will be how to safeguard trustworthy information and journalism within an increasingly AI-driven, platform-dominated media environment. Beyond financial sustainability, key questions remain for policy-makers, researchers and civil society at large: How can media literacy be strengthened at scale? How should content provenance be communicated – through labelling or other mechanisms? And how can issues such as platform concentration, data licensing and copyright be effectively addressed in the AI era?

Amy Ross

Research Fellow

Reuters Institute for the Study of
Journalism, University of Oxford

Nic Newman

Senior Research Associate

Reuters Institute for the Study of
Journalism, University of Oxford

A resilient information ecosystm relies not only on strong institutions, but also on the vitality and independence of media actors. Supporting local journalism, fact-checking networks and open media environments is essential to building public trust and effectively countering harmful information. However, several evolving dynamics are reshaping the information landscape in ways that undermine these efforts:

• The rise of 'local media deserts' – areas with little or no access to reliable local news sources – or communities where residents face significantly reduced access to the news of local public discourse.¹²

  In these environments, people increasingly rely on social media as a primary source of information.

• Shifting media consumption habits are pushing conversations into private, less visible digital spaces – known as 'bounded social media places' – such as private messaging apps and closed groups. These spaces are often viewed as more trustworthy due to their known audiences, controlled visibility and real‑time synchronicity by facilitating continuous conversations and personalized content sharing, often outside the influence of public‑facing algorithms.¹³

  However, their opacity makes it harder to monitor or challenge the spread of misinformation.

As highlighted in the World Economic Forum's 2025 Global Risks Report, it is becoming increasingly difficult for the public to discern where to turn for trustworthy information. Across 47 countries, only 40% of people say they trust most news, and there are concerns about the risk of misinformation and disinformation over the next two years, especially in high-income countries. It ranks among the top 5 perceived risks in 13 countries including India, Germany and Canada, and in the top 10 in 30 additional countries.¹⁴

The media plays a dual role in today's information landscape – acting both as a vital conduit for trusted information and, at times, a vector amplifying harmful narratives. This complexity is critical to understand in humanitarian contexts where media coverage can shape public perceptions of crises, humanitarian actors and the legitimacy and/or effectiveness of aid efforts. Not all media are the same and these distinctions matter:

• Independent journalism plays a watchdog role, conducting fact‑checking, holding power to account and providing nuanced, evidence‑based reporting.

• State‑aligned media often reflect government interests, particularly during armed conflict or emergencies, shaping narratives that support national agendas.

• Commercial and tabloid outlets may prioritize sensationalism and speed over accuracy, driven by attention‑based revenue models and algorithmic amplification.

• Digital and social media platforms can accelerate the spread of both accurate information and harmful narratives, amplifying reach through network effects, algorithms and user engagement.

Meanwhile, media outlets themselves are under growing pressure. Trends documented in the World Press Freedom Index¹⁵ reveal shrinking press freedoms, alongside rising harassment, censorship and targeted information attacks. Many local and independent outlets face severe economic precarity, leading to newsroom closures and weakened on-the-ground reporting. In parallel, regulatory tools such as licensing requirements or anti-fake news laws have, in some cases, been used to suppress independent journalism, rather than protect the public from disinformation.¹⁶

In humanitarian settings, the media plays a powerful role in shaping narratives around aid operations. Politicized or misleading coverage can erode public trust in humanitarian actors, hinder access or even place humanitarian personnel, volunteers and affected populations at risk. Despite these challenges, there are also opportunities for constructive engagement. Partnerships with trusted local journalists can enhance the accuracy, relevance and reach of communication on humanitarian crises. At the same time, media literacy initiatives can help communities distinguish fact from fiction, better understand the risk of harmful information and engage with media content more critically.

Given the diversity of today's media landscape, it is crucial to identify where harmful information is spreading – whether through television, radio, print, digital platforms or social media – and to understand how different communities access and trust information. This requires disaggregated analysis by age, gender, language, geography and other demographic factors, alongside close engagement with community leaders and local stakeholders. As media consumption habits continue to evolve, the need to promote high-quality journalism and empower individuals to navigate an increasingly complex and polarized media environment has never been greater. For the humanitarian sector, this means actively engaging to build understanding of what humanitarian action is and is not, and of the principles that underpin it.

5.6 A threat to humanitarian action and to humanity itself

Harmful information undermines trust, deepens social divisions and weakens the ability of institutions – humanitarian or otherwise – to respond effectively in times of crisis. Addressing it requires a whole-of-society response, not just a humanitarian one. It is intrinsically linked to preserving humanitarian space and the urgent need to prevent the instrumentalization of humanitarian action and actors. Harmful narratives threaten the perceived legitimacy of humanitarian organizations, shaping public perceptions in ways that can directly hinder their ability to access and respond to populations in need. In some cases, this involves misrepresenting an organization's neutrality, intent or operations to incite hostility and disrupt response efforts.

Recognition of this threat is growing. In a significant step, the UN Security Council adopted Resolution 2730¹⁷ in May 2024, spearheaded by the Government of Switzerland and co-sponsored by 98 member states. The resolution explicitly condemns "disinformation, information manipulation, and incitement to violence" against humanitarian and UN personnel. It further raises concern over the increasing use of malicious information and communication technologies including data breaches and information operations that target humanitarian organizations, disrupt their relief operations, undermine trust and threaten the safety and security of their personnel, premises and assets. The resolution encourages member states and the UN system to take appropriate action to address the increasing threat of disinformation campaigns and misinformation that undermine trust in humanitarian organizations, put personnel at risk and hinder humanitarian activities. Importantly, the resolution reaffirms the need for all parties to armed conflict to preserve the ability of humanitarian organizations to act in a manner consistent with the principles of humanity, neutrality, impartiality and independence. This is vital for delivering assistance to persons in need and for ensuring their protection and safety, as well as that of humanitarian personnel.

The UN General Assembly's resolution on 'Countering disinformation for the promotion and protection of human rights and fundamental freedoms' calls on states to counter all forms of disinformation through policy-based measures, including public education, digital literacy and capacity-building initiatives.¹⁸ As highlighted in Chapter 4, on page 157, the Pandemic Agreement is a landmark international treaty aimed at setting global standards for pandemic prevention, preparedness and response. It underscores the importance of trust, transparency and timely information sharing in effective pandemic communication.¹⁹

5.7 Red Cross and Red Crescent Appeal to States

Since the Statutory Meetings of the International Red Cross and Red Crescent Movement began in 1867, themes related to information have rarely been addressed.²⁰ However, 2024 marked a turning point with the adoption of an Appeal to States to take action against harmful information and dehumanizing rhetoric.

Appeal to States:

The appeal highlights the increasing impediments faced by principled humanitarian actors in delivering protection and assistance to people in need including the spread of misinformation and disinformation that imperils humanitarian workers and people in their care.²¹ While previous resolutions have reflected the Movement's longstanding commitments to provide quality information and combat discrimination, the 2024 appeal reflects a heightened sense of urgency. This shift underscores the importance of the information environment, while reinforcing the need to safeguard unhindered and safe access to people in need to carry out principled humanitarian action.

This resolution (appeal) recognizes that harmful information and dehumanizing rhetoric pose serious risks to people in humanitarian need and humanitarian personnel. Addressing this is thus integral to the protection of people and upholding humanitarian operations.

In disasters, crises and other emergencies, laws, policies and plans that promote the availability and integrity of information – and address the challenges of harmful information – are essential to enable and facilitate effective humanitarian action. Harmful information that obstructs humanitarian response, fuels panic and erodes trust runs counter to these objectives. By integrating measures to counter harmful information, states can uphold their responsibilities while protecting the space for neutral, impartial and independent humanitarian action.

5.7.1 The evolving place of information in humanitarian standards

The 1994 Code of Conduct for the International Red Cross and Red Crescent Movement and NGOs in Disaster Relief makes little reference to information, except in relation to the availability of data "pertinent to the implementation of effective disaster response." Building on this, the 1997 Humanitarian Charter highlighted the "transparency of information and decision-making" as essential to effective and accountable humanitarian action. Subsequent frameworks have placed growing emphasis on information as central to dignity, participation and protection.

The Humanitarian Charter and Sphere Standards (1997–present) identify access to information as a core element of participation; the Core Humanitarian Standard on Quality and Accountability (2014) includes explicit commitments to information-sharing and communication with affected people; and both the ICRC Rules on Personal Data Protection and IFRC Policy on the Protection of Personal Data frame personal data and information management as integral to protection, dignity and legal compliance. In parallel, the CDAC Network Guidelines advanced the principle of 'communication as aid,' focusing on community engagement and misinformation management, while the International Civil Society Centre's Solidarity Action Network has more recently turned attention to the impact of disinformation campaigns on civil society.

Taken together, these instruments underscore that information is not peripheral to humanitarian action but fundamental to its effectiveness, accountability and legitimacy. Looking ahead, this recognition should also encompass harmful information, ensuring that standards explicitly address how it can undermine safety, dignity, humanitarian access and action response.

5.7.2 The tech sector: Platforms, power and responsibility

The explosive growth of online content, driven by new technologies and user-generated platforms, has made harmful information more prevalent and harder to detect, moderate and remove. What began as spaces for personal connection have evolved into powerful tools for influence, distortion and manipulation, shaped by the speed, scale and reach of digital information. Social media platforms are designed to reward engagement, not accuracy. What goes viral is not necessarily what is true. Built around attention-based algorithms, these platforms fuel contests over narrative and influence with real-world consequences. These platforms are designed for maximum engagement, where every like, share and comment release a hit of dopamine, reinforcing addictive use and deepening echo chambers.²⁸

The rise of AI has further concentrated power in the hands of a small number of companies that develop and control the systems including algorithms shaping what billions of people see, search for and share. Their influence extends beyond access – these systems increasingly affect how public opinion is formed, how issues are framed and which narratives gain prominence. Generative AI has drastically lowered the barriers to creating and distributing false or misleading content, across text, video, audio and images. Distinguishing AI-generated from human-made content is increasingly difficult. While some material is simply the result of error or AI hallucination, other content is deliberately produced by threat actors – state actors, activist groups or individuals – aiming to mislead, influence or disrupt. This rapid, low-cost production and deployment of synthetic content presents significant challenges for information integrity.

5.8 UN action on AI and information integrity

The UN General Assembly's March 2024 Resolution on AI systems²⁹ highlights the growing risks to information integrity, access to information and human rights. It warns that the improper or malicious design, development, deployment and use of AI systems could undermine the Sustainable Development Goals (SDGs), deepen digital divides and reinforce structural inequalities and biases. The resolution raises concerns about the potential accidents and the risk of compounded threats from malicious actors. It encourages member states and stakeholders to develop effective and interoperable tools, reliable content authentication and provenance mechanisms – such as watermarking or labelling. These measures would enable users to identify information manipulation, determine the origins of and distinguish between authentic and AI-generated or manipulated digital content. It further encourages efforts to strengthen media and information literacy to strengthen societal resilience.³⁰

The June 2023 UN policy brief on Information Integrity on Digital Platforms was the result of consultations developed under the auspices of the Secretary-General as part of the Our Common Agenda and Summit of the Future initiatives. This seeks to provide a concerted global response to information threats in the information environment. The principles aim to guide member states, digital platforms and other stakeholders in fostering a more inclusive, transparent and safe digital space, particularly freedom of opinion, expression and access to information.³¹ In September 2025, the UN released the first in a new Issue Brief series entitled From Principles to Practice: Strengthening Information Integrity.

The UN's 2024 Governing AI for Humanity report³² warns that large parts of the world are excluded from international AI governance conversations, with only seven countries party to seven prominent non-UN AI initiatives and 118 countries not party to any.³³

5.9 Content moderation and the power of platforms

Content moderation has become one of the most powerful and contested functions exercised by global technology platforms. Companies behind social media and messaging services play a central role in shaping the information ecosystem, determining what is amplified, removed or monetized. The model of social media platforms – functioning as intermediaries allowing user-generated content to spread at scale – has effectively shifted much of the responsibility for identifying and managing harmful information onto the private sector. This places content moderation and governance in the hands of technology companies rather than public institutions. Content moderation is carried out under each platform's own terms of services and definitions, typically through a combination of human moderators and automated systems. Platforms may block or remove content either in response to legal requests or based on internal assessments. However, definitions of harm vary and enforcement is uneven.

This power is exercised with limited transparency, especially regarding how content is moderated and curated. Many governments have been reluctant to regulate legal but harmful information citing their obligations to uphold freedom of expression. At the same time, global technology platforms often comply with national content laws, even when those laws conflict with international human rights standards. This results in an uneven and fragmented landscape, where users' access to content depends largely on their location and the legal environment of that jurisdiction. While some platforms publish transparency reports outlining how often governments request content removal, reporting practices vary widely and often lack the detail necessary for meaningful public accountability.

While content regulation can be lawful and necessary, especially to prevent harm, some states use it to control narratives, limit transparency and suppress dissent. This raises a deeper set of questions: Who decides what constitutes harmful information? Who is accountable? And how can we safeguard fundamental rights in a fragmented and global information space?

Under international law, for example, Article 19 of the International Covenant on Civil and Political Rights, freedom of expression may only be restricted when three conditions are met: the restriction must be provided by law (clearly defined in legislation), necessary (to achieve a legitimate aim) and proportionate (not excessive or overly broad). Yet many current censorship practices fail to meet these criteria, raising serious concerns about compliance with international standards.

The UN Guiding Principles on Business and Human Rights and related standards provide important guidance for technology companies. These principles call for heightened human rights due diligence and risk management to prevent and address adverse impacts, including from the spread of harmful information. This is particularly critical in contexts affected by humanitarian crises, where such information can exacerbate vulnerabilities and fuel harm. Companies are expected to invest in and implement robust measures to mitigate potential negative impacts of their activities on people's safety and dignity. This includes ensuring that their policies, procedures and practices are consistent with international human rights standards and take into account international humanitarian law.

Complementing this, the UNESCO framework for regulating digital platforms is grounded in the recognition that information is a public good. As such, it calls for government action to protect and support the integrity, accessibility and equitable availability of information ecosystems.³⁴

Meta (formerly Facebook), for example, does not explicitly list humanitarian organizations as targets in its reporting on inauthentic behaviour. Yet, its findings on narrative manipulation and inauthentic behaviour in contexts of armed conflict have clear implications for humanitarian actors, whose operations may be disrupted or delegitimized by such tactics.³⁵

Microsoft highlights that not all threats posed by AI originate within the systems themselves; many stem from the broader information ecosystems.³⁶ Two prominent categories stand out:

• Impersonation: AI‑generated deepfakes (audio, video or images – see Annex I: Glossary, on page 353) increasingly enable impersonation of individuals, with serious risks including fraud, blackmail, coercion, defamation and information warfare.

• Content production: AI tools can be misused to produce harmful synthetic content at scale, including disinformation, spam, non‑consensual intimate imagery and grooming scripts. These threats are often amplified versions of pre‑existing problems, now rendered faster and more pervasive by generative AI.

Microsoft has also called on states to adopt clear norms and restrictions to curb harmful foreign influence operations, particularly those targeting crisis and emergency contexts, humanitarian and emergency response organizations, elections, marginalized communities and protected groups such as ethnic minorities and LGBTQ+ populations. It advocates for limits on the use of certain tools and techniques including synthetic media (e.g., deepfakes) and emphasizes that social media data from foreign citizens should not be exploited for influence operations.³⁷

5.10 From self‑regulation to state oversight: The evolving governance of online content

Despite their language of community, digital and social media platforms are, above all, businesses. Their primary responsibility is to shareholders, not the public good – so the metrics that matter most are financial. This profit-driven logic shapes how platforms approach both problems and solutions. With an engineering-first mindset, they often frame even complex societal challenges as design problems solvable through technical fixes. Ultimately, platform terms of service – written by private companies – administer communities of unprecedented scale.³⁸ Yet fundamental dilemmas remain: Should these companies restrict the flow of information? What content should be restricted, how and by whose standards?

The US legal framework, particularly Section 230 of the Communications Decency Act, has played a central role in shaping this landscape.³⁹ It shields US platforms from liability for user-generated content while also protecting them when they moderate content in "good faith." This approach effectively left much of the internet to self-regulate, with content moderation emerging as a corporate practice – not out of legal obligation, but as a strategy to pre-empt stronger regulation. In contrast, enforcement around intellectual property – governed by the more stringent 1998 Digital Millennium Copyright Act has drawn much clearer lines.⁴⁰

The European Union's Digital Services Act, which came fully into force in 2024, establishes a comprehensive legal framework to regulate online services and address the spread of illegal and harmful content, including misinformation and disinformation, while safeguarding fundamental rights such as freedom of expression. Non-compliance can result in significant penalties, including fines of up to 6% of a company's global annual turnover. The Digital Services Act seeks to strike a balance between mitigating systemic societal risks associated with online content and protecting freedom of expression and information.⁴¹ Key provisions relevant to harmful information include:

1. Systemic risk management: 'Very large online platforms' (with more than 45 million active users in the EU) and 'very large online search engines' are required to identify, analyse and mitigate systemic risks stemming from the design, functioning and use of their services. This includes risks posed by disinformation that could threaten democratic processes, public health or public security.

2. Transparency and accountability: Platforms must disclose their content moderation policies, advertising practices and key parametres of algorithmic decision‑making and provide users with accessible complaint‑handling and appeal mechanisms.

3. Code of practice: In February 2025, the strengthened EU Code of Practice on Disinformation was formally integrated into the Digital Services Act framework. Measures include demonetizing disinformation, strengthening fact‑checking and improving access to reliable and authoritative information.

4. Enhanced user‑reporting mechanisms: Online platforms must provide accessible and effective systems for users to report illegal content.

5. Crisis‑response mechanism: In exceptional circumstances, such as threats to public security or public health, the European Commission may require platforms to take targeted measures to address the rapid dissemination of harmful information during crisis periods.⁴²

These developments build on earlier EU initiatives, including the European Commission's Code of Practice on Disinformation and the Strengthened Code of Practice on Disinformation.⁴³

5.11 Framing a response: Supply and demand solutions to disinformation

Professor Anya Schiffrin of Columbia University developed a 'taxonomy of solutions' – an analytical framework that categorizes various policy responses to online misinformation and disinformation into two broad approaches: supply side and demand side.

5.12 Civic trust and societal resilience

So what I would want to suggest is: they should start with the ground level in schools. Educating people in schools, telling them the quality source of information. Once that is done, those children will keep that and they will grow with that kind of mentality. So the education part should continue using the Ministry of Education in schools that should know the correct source of information. Our community meetings like the church and other political gatherings should also be emphasized. But now the only problem is because of these issues, political issues, people want to gain support, they want to use whatever means. But what I would suggest is let the children be taught the correct things from the Ministry of Education first."

Some countries have adopted whole-of-society approaches to build societal resilience – to inoculate – against information threats. These include education programmes, public tracking of foreign disinformation, election protection measures, transparency for political advertising and campaign activities. In addition, some governments use legal measures, such as laws for the removal of illegal content.

International research⁴⁷ highlights several systemic challenges facing civil society actors, which hamper their ability to assess information threats and develop effective responses. While a range of initiatives are emerging, many remain isolated, pointing to an urgent need for stronger coordination, knowledge-sharing and the importance of skill diffusion to prevent duplication and build collective capacity. A critical constraint is the limited access to social media data, which limits civil society's ability to assess the scale of disinformation and to evaluate the effectiveness of different interventions. Furthermore, engagement with technology platforms remains uneven: many organizations, particularly in the Global South, struggle to have their concerns heard and are frequently overlooked by platform decision-makers.

The OECD identifies five key determinants of public trust in government⁴⁸ that may provide some interesting insights for civil society in building institutional trust:

• responsiveness and reliability in delivering services and anticipating needs (reflecting competence)

• perceptions of integrity, openness and fairness (reflecting public values).

Academic literature further distinguishes between: trust in competence – the ability to deliver on expectations – and trust in intentions – the perception that actions are taken in good faith. These two dimensions are interdependent and influence how people evaluate institutions. Trust is ultimately built on both performance and principle – the perceived capability of institutions and the values guiding their actions. Transparency, fairness and accountability serve as mutually reinforcing attributes that strengthen both pillars and are essential for building societal resilience in the face of harmful information.

Concluding remarks: A collective responsibility for preserving principled humanitarian action

For the humanitarian sector, strengthening cross-sector engagement is essential – not only to ensure the flow of trustworthy information but also to identify and respond to harmful information. Disruption alone is not enough; responding effectively requires collective action, similar to the collaborative efforts developed for global health security during the COVID-19 pandemic.

Harmful information is borderless, adaptive and often directed at the people most at risk. Algorithmic systems amplify its spread, while global disparities in data governance, AI capacity and political and societal polarization intensify its reach. Effective responses require greater algorithmic transparency on how content is sorted, ranked, amplified and targeted, combined with approaches that protect privacy, dignity and authenticity.

Resilience must be built at every level – individual, institutional and societal – drawing on community trust, local knowledge, behavioural science, digital and information literacy. Evidence shows that simple interventions, such as digital prompts,⁴⁹ can reduce the sharing of false content, underscoring the value of pairing regulation with context-specific solutions. Governments, platforms, civil society and researchers must work together to test and scale what works. This underscores the importance of pairing regulation with behavioural science and practical interventions that are context specific. Governments, platforms, civil society and researchers must work together to test what works, for whom and in which context.⁵⁰

Humanitarian crises – from pandemics to armed conflict – create fertile ground for harmful information. Information related to the COVID-19 pandemic exposed both the speed of harmful information spread and the cost of delayed action. Meeting the challenges of identifying, preventing and mitigating harmful information requires a systems approach and urgent, cross-sectorial action. Fragmentation between humanitarian actors, digital rights advocates and the technology communities must be bridged. This includes creating space for emerging questions, fostering diverse and underrepresented perspectives, and prioritizing principles and accountability. Strong, forward-looking partnerships are essential, not only to shape governance frameworks, but to influence the design and deployment of tools used to navigate this rapidly evolving landscape.

Ultimately, states must act. As underscored in the Movement's Appeal to States, addressing harmful information is now central to principled humanitarian action. Trusted information may not be water, food or shelter but it is imperative to accessing all three, as well as to ensuring safety, dignity and autonomy.

Asks, aims and recommendations

Endnotes