Archive · 29 Jul 2025
Humans Struggle to Spot AI-Generated Images, Study Finds
Imagine a game where you have to tell if a picture is real or made by a computer.
Scientists used this game to test how good people are at spotting fake pictures created by AI.
They found that people aren't very good at it, only a little better than guessing with a coin.
The study used pictures made by popular AI tools.
This means AI is getting really good at making pictures that look real, and it's becoming harder for us to tell the difference.
Researchers studied human ability to identify AI-generated images.
Participants played an online game called "Real or Not Quiz".
The study used images generated by AI tools.
The study aimed to measure how well humans can distinguish real vs. AI images.
Humans are only slightly better than a coin toss at spotting AI-generated images.
- Who
- Researchers from the Microsoft AI for Good Lab conducted the study.
- What
- Study on human ability to identify AI-generated images.
- When
- Not specified in the article.
- Where
- The study used an online game.
- Why
- To measure how well humans can distinguish between real and AI-generated images.
- Study Focus
- Distinguishing AI-generated images
- Method
- Online game "Real or Not Quiz"
- AI Tools Used
- ChatGPT, Gemini, and Claude
- Main Finding
- Humans slightly better than coin toss