US EditionPublished • loading... • Updated
AI's Blind Spot: Tools Fail to Detect Their Own Fakes
AI chatbots often misidentify AI-made images as real, undermining fact-checking amid reduced human oversight and increasing user reliance, AFP said.
- This year, AFP found AI chatbots failed to spot fabricated photos, including a fake image of Elizaldy Co and a manufactured protest photo from Pakistan-administered Kashmir.
- Experts note the root cause is large models mimic patterns without specialised visual-forensics, and the problem grows as Meta ended third-party fact-checking earlier this year.
- Columbia University's Tow Center for Digital Journalism tested seven AI chatbots on 10 images earlier this year, and all failed to identify provenance while AFP traced the Co image to Google's Gemini and Nano Banana, created by a Philippine web developer.
- Internet users are increasingly turning to chatbots to verify images, but the fabricated Elizaldy Co photo garnered over a million views, illustrating real-world misinformation spread.
- Experts warn AI verification modes can assist but not replace trained human fact-checkers, with Rossine Fallorina stressing, it's "We can't rely on AI tools to combat AI in the long run.
Insights by Ground AI
17 Articles
17 Articles
Test your knowledge of AI hallucinations, digital learning materials and disinformation
·Vienna, Austria
Read Full Article
AI's blind spot: tools fail to detect their own fakes
Internet users are increasingly turning to chatbots to verify images in real time, but the tools often fail, raising questions about their visual debunking capabilities at a time when major tech platforms are scaling back human fact-checking.
·Malaysia
Read Full Article
When AI Tries To Identify AI: Chatbots Fail To Detect Images They Created
When outraged Filipinos turned to an AI-powered chatbot to verify a viral photograph of a lawmaker embroiled in a corruption scandal, the tool failed to detect that it was fabricated -- even though it had generated the image itself.
·New Delhi, India
Read Full Article
Coverage Details
Total News Sources17
Leaning Left2Leaning Right5Center2Last UpdatedBias Distribution56% Right
Bias Distribution
- 56% of the sources lean Right
56% Right
L 22%
C 22%
R 56%
Factuality
To view factuality data please Upgrade to Premium
