A new study reports that some AI-generated faces can seem more trustworthy than real human faces, making it difficult for people to reliably tell which images are AI-made. According to coverage of the findings, researchers tested participants’ ability to distinguish between faces generated by artificial intelligence and photographs of real people, and found that accuracy is limited. The reporting also notes that AI-generated images may exploit visual cues that influence impressions of trustworthiness, leading to higher trust ratings for certain synthetic faces. While the study’s results point to a potential risk for perception and decision-making, the articles do not provide extensive detail on methods, sample size, or how “trust” was measured. The coverage encourages readers to take a test or attempt to identify which faces are AI-generated, reflecting the study’s emphasis on the challenge of visual detection. Overall, the articles converge on the central message: AI face synthesis can affect judgments of credibility and may not be easily distinguishable from real faces.