AI could fuel wave of Holocaust denial, UNESCO finds

A United Nations agency is warning that developments in artificial intelligence (AI) could fuel a new surge in Holocaust denial.

A UNESCO report published on Tuesday suggests that AI systems could misrepresent information about the Holocaust because they could be trained to answer questions about it from denial websites.

It said that AI has been used to “distort Holocaust-related content, creating fabricated testimonies, and even altering historical records”. UNESCO also warned about deepfakes – realistic images or videos that suggest the Holocaust didn’t happen or was exaggerated.

One application already lets users speak directly to prominent Nazi figures like Adolf Hitler.

Generative AI models also invent or “hallucinate” events, personalities and history when they don’t have access to enough data.

Google’s Bard and ChatGPT have produced content about the Holocaust-related events “which never took place,” the report continued.

ChatGPT fabricated the concept that Nazis drowned Jews in rivers and lakes in a so-called ‘Holocaust by drowning’ campaign. Bard, on the other hand, generated fake quotes from witnesses to support fake narratives about the massacres.

“If we allow the horrific facts of the Holocaust to be diluted, distorted or falsified through the irresponsible use of AI, we risk the explosive spread of antisemitism and the gradual diminution of our understanding about the causes and consequences of these atrocities,” Audrey Azoulay, UNESCO’s director general, said in a statement accompanying the report.

UNESCO’s report called on tech companies to establish ethical rules for the development and use of AI, to reduce the chances of unreliable information and to prevent bad actors from harnessing their programmes to encourage violence and spread lies about the Holocaust.

The report was published in partnership with the World Jewish Congress.