The United Nations Refugee Agency (UNHCR) has issued an urgent warning that the rapid proliferation of misinformation, hate speech, and AI-generated deepfakes is severely undermining the safety of displaced populations and humanitarian personnel worldwide. Speaking at the recent AI for Good Summit in Geneva, the agency underscored that while artificial intelligence presents significant potential to streamline humanitarian support, its capacity to accelerate the spread of malicious content creates a volatile environment. For refugees, who often rely on digital platforms to navigate their displacement, this shift represents a profound threat to their security and access to essential protection services.
As AI technology continues to evolve, the UNHCR emphasized that refugees and humanitarian organizations must be prioritized in global governance discussions. The agency urged governments, researchers, and technology corporations to collaborate on robust safeguards that prevent AI from being weaponized against vulnerable groups. A critical aspect of this necessity is improving content moderation and ensuring that safety tools are linguistically inclusive. Without structural oversight that considers those on the margins of society, the agency warns that new digital systems risk exposing refugees to heightened levels of harm rather than providing the intended benefits of technological advancement.
The consequences of this digital crisis are far from contained within online forums; they have bled into the physical world with devastating effects. Misinformation and dehumanizing digital narratives have been directly linked to real-world violence, restricted access to employment and education, and, in extreme cases, the loss of life and forced displacement. The UNHCR pointed to alarming incidents in Libya, where AI-generated deepfakes of agency representatives and the leaking of sensitive staff locations have actively compromised humanitarian operations and placed aid workers in immediate danger. For communities like the Rohingya, this online hostility—which often persists long after they have sought refuge elsewhere—severely hinders their ability to integrate into host societies.
Data from a recent UNHCR survey highlights the scale of this issue, revealing that 93 percent of humanitarian staff have witnessed how misinformation and hate speech directly impede their work. The burden of this digital abuse is not shared equally, as women—both refugees and aid workers—are disproportionately targeted by harassment. Generative AI has lowered the barrier to entry for producing abusive content, allowing bad actors to scale their campaigns with alarming efficiency. The agency noted that these tactics are frequently utilized by smugglers and human traffickers, who deploy deceptive AI-driven content to lure desperate individuals with fake promises of safe migration, jobs, or legal status.
In response, the UNHCR is championing a global effort to prioritize information integrity as a core pillar of international refugee protection. This initiative includes the establishment of the Community of Practice on Information Integrity in Humanitarian Contexts, supported by Switzerland, which fosters collaboration between civil society, governments, and the private sector. Furthermore, the agency has deployed an Information Integrity Response Toolkit, already utilized by approximately 2,000 practitioners, to help local teams assess and mitigate digital risks in real time. The agency categorically stated that these efforts are not intended to impede freedom of expression, but rather to protect life by curbing the deceptive narratives that facilitate exploitation and violence.
Ultimately, the UNHCR maintains that the preservation of trusted information is no longer a secondary concern, but an essential component of modern humanitarian protection. As international frameworks for AI governance continue to take shape, the agency is calling for humanitarian considerations to remain at the center of the debate. By ensuring that technology companies and nations adopt a human-centric approach to AI, the international community can begin to dismantle the digital threats that currently endanger the world’s most vulnerable, fostering an environment where technology serves to assist and protect rather than endanger and deceive.
