In the world flooded with AI images, it’s essential that we can detect them fast. OpenAI has announced that it’s releasing a tool that will help in this process. According to the company, it will be able to detect AI-generated images with a 99% accuracy.
The announcement came during the Wall Street Journal‘s Tech Live conference on Monday. Mira Murati, OpenAI’s Chief Technology Officer, shared that the tool is currently undergoing internal testing. There’s no word on the public release date, but Murati noted that the tool is “99% reliable” already. Murati added that she wanted the tool to be designed so that OpenAI’s users “don’t feel monitored.”
Can AI detection tools be 100% accurate?
It’s worth noting that AI image detection tools are not new to the market. The most popular one is probably AI or Not, but it’s not unmistakable. As PetaPixel [sensitive topic] recently reported, this tool marked a photo from the Israel-Hamas war as AI-generated. However, an expert confirmed that it was, in fact, a real photo. It appears that people don’t only use AI-generated images to spread fake news – but they also use AI detectors to discredit the horrors of war.
For now, there is no AI detection tool that is 100% accurate. Still, they can at least help you figure out whether Giants of Kandahar or people kicking alligators are real. As for more serious stories – please rely on your common sense and don’t believe everything you see.