OpenAI launches ChatGPT for Teens amid child safety expert skepticism

OpenAI released an age-gated version of its chatbot called ChatGPT for Teens, which routes users predicted to be under eighteen to a default experience that includes educational features such as Study Mode and data visualizations. The release followed a lawsuit alleging that the original chatbot contributed to a teenager’s suicide. OpenAI also announced expanded safety notifications that would alert parents linked to accounts when unsafe discussions about eating disorders are detected.

Child safety experts interviewed by Engadget expressed skepticism, saying the company must prove its safety commitments before the product can be trusted. They highlighted the lack of published data on the age-estimation system’s false-positive and false-negative rates, noting that the automated system does not detect nearly all teen users. Experts also warned that teens could try to bypass the age-gating using tools like VPNs or other workarounds.

A central promise from OpenAI is that all flagged content will be reviewed by full-time employees and that parents will be notified within an hour. Experts questioned whether the automated classifiers that flag content for human review are reliable enough to justify this workflow. They also pointed out the absence of any accountability mechanism if the system fails to protect minors.

The experts cited past experience with social media platforms where trusting companies to self-police did not work, and they noted that OpenAI’s earlier suicide-related content restrictions had eroded over time. They argued that without independent testing and transparent reporting, the safety features remain unproven. The need for verifiable metrics was a recurring theme in their comments.

For engineers building or deploying similar AI services, the episode shows that age-gating must be accompanied by measurable error rates and mechanisms to detect evasion. It also shows that content-flagging pipelines require demonstrable human-review capacity and clear escalation paths. Finally, transparent safety documentation and third-party validation become prerequisites before releasing products aimed at minors.