The rapid expansion of artificial intelligence has introduced a dangerous new front in the humanitarian landscape, as the UNHCR recently warned that AI-driven misinformation, hate speech, and manipulated content are actively endangering refugees and the aid workers who support them. While digital tools have long been used to connect people, the rise of powerful generative AI has made the creation of high-quality “deepfakes” and coordinated disinformation campaigns both cheaper and more accessible. These tools are no longer just an abstract concern for policymakers; they are becoming a direct threat to the lives and safety of the world’s most vulnerable populations, turning the digital world into a source of real-world harm.
The scale of this issue is profound, as highlighted by a recent survey of UNHCR personnel, where a staggering 93 percent of staff confirmed that their work on the ground has been hampered by viral falsehoods and toxic rhetoric. It is one thing to see a celebrity in a deepfake video, but it is entirely different when AI is used to manufacture fake warnings, impersonate agency leaders, or spread dehumanizing propaganda against displaced people. This digital pollution is infecting the communities that host refugees, turning neighbors against one another and poisoning the atmosphere in which aid organizations must operate. When falsehoods move from viral posts to the streets, the results are often tragic, as misinformation has been directly linked to riots, physical assaults, and even instances of deadly violence.
The societal impact of this digital distortion is multifaceted, effectively acting as a barrier to safety and dignity for those who have already lost everything. When someone fleeing conflict arrives in a new environment, they depend on accurate information to navigate their rights, find legal employment, and access education; misinformation robs them of this agency, trapping them in a state of confusion and fear. By fueling xenophobia and scapegoating, the hateful narratives amplified by AI make the process of social integration nearly impossible. In many cases, these rumors dehumanize displaced individuals, stripping them of their humanity in the eyes of their host communities and justifying further mistreatment or even forcing them into secondary displacement.
A chilling example of this reality can be seen in Libya, where the weaponization of online information has become a grave threat. In recent months, humanitarian efforts have been severely disrupted by malicious actors using AI to generate fake videos that impersonate senior UN officials or depict local aid workers as traitors. These tactics are calculated and cruel, often designed to incite local populations to identify and target relief centers or reveal the secret locations of staff members. By eroding trust and creating a climate of constant suspicion, these bad actors are effectively cutting off the lifeline to refugees, making it increasingly dangerous for aid agencies to provide even the most basic services.
Beyond the threat of direct violence, criminals and exploiters are increasingly using technological deception to prey on the desperation of migrants. Sophisticated groups are leveraging digital platforms to craft elaborate lies, dangling false promises of safe passage, legal status, or non-existent job opportunities in front of people who are often at their lowest point. These are not merely digital scams; they are conduits to exploitation, steering vulnerable families directly into the hands of human traffickers and smugglers. By manipulating the hopes of the displaced, these criminal networks turn AI-driven digital influence into a high-tech dragnet that leaves refugees more exposed and financially destitute than before.
In response to this crisis, the UNHCR is making an urgent, human-centered appeal to the global community during the AI for Good Summit in Geneva. They are calling upon governments, tech giants, and data scientists to move beyond profit and technical benchmarks to ensure that humanitarian protections are baked into the core architecture of future AI systems. This includes much smarter content moderation that can understand the nuance of humanitarian crises and function effectively in lesser-spoken, regional languages. Ultimately, the agency is demanding a shift in focus: leaders must stop prioritizing speed and scale and start placing people at the center of safety protocols, ensuring that human rights and social cohesion are protected against the destabilizing forces of the digital age.