Humans are really bad at detecting fake AI images, study shows
The study conducted by Microsoft's AI for Good Lab found that humans are surprisingly poor at distinguishing AI-generated images from real ones. The research involved over 12,500 participants who analyzed approximately 287,000 images, with a success rate of only around 62% in identifying real and fake images. The study showed that it is slightly easier to identify fake images of faces compared to landscapes, but the difference is marginal. In response, Microsoft is advocating for clearer labeling of AI-generated images, though critics argue that this can be easily circumvented by cropping the images. The findings highlight the growing challenge in detecting AI-generated content, which has significant implications for the future of visual media and online information.
Note: This is an AI-generated summary of the original article. For the full story, please visit the source link below.
