Focus area
Information Integrity
We monitor misinformation and disinformation and listen to the concerns behind them, so that teams understand what people ask and fear, and respond to what matters most.
CHALLENGES
Misinformation and disinformation now move faster and in more languages than any team can follow, and AI has made false content cheaper to produce. Most investment has gone into debunking. That work matters, and it works better when it starts from understanding why a narrative took hold, because behind a false claim there is usually a question nobody answered or a fear nobody acknowledged.
Why it matters
Left unaddressed, false information discourages people from seeking help, puts staff and operations at risk, and wears down trust in the organisations that serve them. Without an understanding of what lies behind it, responses stay short term and the distance between institutions and communities grows. A late response is of little use, so teams need to see early which narratives deserve attention and which can be left alone. They also need what they learn to shape real decisions, whether on messaging, programming or budget.
How we work
Seeing what circulates
Which narratives spread, in which languages and on which channels, and whether they are shared in good faith or built to mislead, because the two call for different responses.
Seeing it early
ur Misinformation Virality Detector, built on research from the University of Zurich, analyses how content is constructed instead of what it says, so it catches narratives because they are new and estimates how far each will travel. It does not rule on what is true, and it supports the fact checkers already at work.
Listening to communities
Hotlines, surveys and field channels show what people ask, fear and need, including what never reaches social media.
Designing and testing the response
Findings become Q&A, key messages and escalation protocols, and we test messages with the audiences concerned through AI Personas before anything goes out.
Building internal capability
Training on information integrity, media and AI literacy, prebunking and crisis simulation, plus the evaluation frameworks that show whether it worked. All sources are analysed around your questions and validated by analysts, with each dataset visible only to its owners.
Our work in practice
Rumours and misinformation in health emergencies
We follow what communities say and ask during health emergencies, so that governments and partners have fast evidence on rumours and concerns while an outbreak is live. In Eastern and Southern Africa, our listening reports during cholera, mpox and Ebola outbreaks give UNICEF offices and their partners that evidence as the situation develops. In Mozambique, community feedback reporting moved from a monthly to a weekly cycle, so that the office can respond while a concern or rumour is still live.
Using community feedback to detect online harms
We use social and community listening to detect the rumours and harms that affect families, so that campaigns address what people actually experience. For UNICEF Pakistan, we analysed more than 1.9 million public mentions to guide a child online protection campaign, which brought out concerns about cyber harassment and the social pressure that discourages reporting, as well as catfishing and deepfakes used to deceive or exploit girls in particular. The campaign was reoriented around online harassment, catfishing and positive parenting, with short videos in regional languages shared through community influencers. The office uses the same listening for its vaccination campaigns, where rumours about safety, fertility and religion are identified and answered with actions tailored by province and audience.
Hostile narratives and misinformation in displacement crises
We track how displaced people are discussed, so that teams can respond both to hostile narratives and to the questions host communities are asking. Through the war in Ukraine, we analysed how refugees were being discussed across Europe and Central Asia, following misinformation directed at them alongside those questions. We worked the same way during the Ceuta crisis, detecting misinformation as it appeared around the crossings from Morocco into Spain and following the situation through i-Hear-U, the platform we built for UNICEF.