AI or Not misidentifies Hamas baby victim as deepfake
AI image detector AI or Not has come under fire for failing to identify a photograph of a baby apparently burned alive by Hamas, raising questions about the tool's accuracy and marketing claims, and fueling doubts about Israel's communications integrity. California-based Optic's AI or Not tool identified a photograph of what Israel said was a burnt corpse of a baby killed in Hamas’s 2023 attack on Israel as being generated by AI. Its results were tweeted as genuine by US Jewish conservative pundit Ben Shapiro to his hundreds of thousands of followers, prompting people to suggest the Israeli government had deliberately been spreading disinformation. Optic claims '95% accuracy' on its website. But deepfake experts such as UC Berkeley professor Henry Farid said the 'Hamas' image showed no sign of being a deepfake. Earlier in 2023, a test carried out by Bellingcat concluded that AI or Not performed 'impressively well' on high-quality, large, and watermarked AI images, but was 'unimpressive' when it came to compressed AI images. Deepfake Deepfakes (a portmanteau of ' deep learning ' and 'fake') are images, videos, or audio which are edited or generated using artificial intelligence tools, and which may depict real or non-existent people. They are a type of synthetic media. Source: Wikipedia 🔗 System 🤖 AI or Not 🔗 Operator: Ben Shapiro Developer: Optic Country: Israel; Palestine; USA Sector: Govt - military; Politics Technology: Machine learnin g Issue: Accuracy/reliability; Mis/disinformation; Transparency I nvestigations, assessments, audits 🧐 Bellingcat (2023). https://www.bellingcat.com/resources/2023/09/11/testing-ai-or-not-how-well-does-an-ai-image-detector-do-its-job/ New York Times (2023). How Easy Is It to Fool A.I.-Detection Tools?
- Date it happened
- 2023-10-01
- Organisation involved
- Ben Shapiro
- Product, system or model
- AI or Not
This incident was imported from AIAAIC and is used under CC BY-SA 4.0. Our additions to it — the structured fields, the translation, the checks against other reports — are published under the same licence.
This is a record of what was reported, not a finding that anyone broke the law. If it names your organisation and you believe it is wrong, the corrections process is free and open to everyone.