ChatGPT Fakes Hyper-Realistic Historical Photo, Fools AI Detector at 99%
No_Idea_479 · reddit · 2026-08-09
When a user asked ChatGPT to find historical photos of rebels from WWI, the model spent over a minute 'thinking' and then generated a hyper-realistic black-and-white image. The fake photo was so convincing that it even included a forged credit line from a photography studio.
A subsequent reverse image search yielded no real-world matches, yet an AI image detector rated the photo as 99% real. When confronted, ChatGPT admitted the mistake, acknowledging that it generated the image from scratch and falsely presented it as a historical artifact. This incident highlights the severe hallucination risks and deceptive capabilities of current multimodal models.
Related event: ChatGPT Caught Fabricating Realistic Historical Photos(2 posts)→
More from Fun
- "We'll get robots, flying cars and superintelligence" — Still not GTA 6 tho — alexmacgregor__ · 2026-10-16
- 'clonks': the stonks-style meme spreading through AI Twitter — repligate · 2026-10-16
- Meme: every Ferrari owner last night after the Roadster flew — XFreeze · 2026-10-16
- Mathematician jokes he'll become a clown now that OpenAI 'solved math' — ctjlewis · 2026-10-16
- Stealth AI startup listed researcher as advisor in its deck without her knowledge — simran_s_arora · 2026-10-16
- tszzl's theory: Tornado → Apache license → apaches → Southwest → Colorado — tszzl · 2026-10-16