OpenAI released ChatGPT for Kids this week, limiting certain features via an age recognition system. Child safety experts have raised concerns over the mechanism’s transparency and effectiveness, demanding extensive independent testing before it is recommended to parents and teens. They argue the system lacks accountability, its accuracy is unverified (no public data on false positive or negative rates), and it may fail to block harmful content as seen in previous attempts. Experts also question the reliability of the age recognition tool and OpenAI’s claim that all flagged content is reviewed within an hour, calling for more detailed data on review processes. They suggest automatic protection features should be built into the product design rather than relying on users to manually set up complex controls.