Common Sense Media deems ChatGPT for Teens an unacceptable risk
ai-safety openai
| Source: Mastodon | Original article
Common Sense Media, a youth‑safety nonprofit, labeled OpenAI’s ChatGPT for Teens an unacceptable risk despite its teen‑focused guardrails.
OpenAI’s “ChatGPT for Teens,” launched in August with promises of tighter safety controls, has been labelled an “unacceptable risk” by the nonprofit Common Sense Media. The organization’s Youth AI Safety Institute tested more than 4,000 prompts across the teen‑focused version and found that key safeguards repeatedly failed.
The assessment highlights several shortfalls. Parent‑linked accounts did not generate any notifications during a 60‑minute simulated conversation about self‑harm or suicide, despite OpenAI’s claim that dangerous chats would trigger alerts. Hotline referrals for crisis‑related prompts fell from 33 % to 23 %, and for depression‑related prompts from 63 % to a mere 3 % after the teen mode went live. In addition, the “Study mode” and other parental‑control features were found to be ineffective, and the teen‑specific “friend‑like” behavior restrictions did not activate even when accounts were explicitly set to teen status.
The findings matter because they challenge OpenAI’s narrative that the product is safer for younger users than the standard ChatGPT. With AI assistants increasingly embedded in education and home environments, unreliable safeguards could expose minors to harmful content, misinformation, or manipulative interactions. The report also adds pressure on regulators who are beginning to scrutinise AI safety for children, echoing broader concerns raised in recent coverage of OpenAI’s safety practices.
What to watch next includes OpenAI’s response – whether it will roll out software patches, redesign parental‑notification mechanisms, or pause the teen offering altogether. Industry observers will also be tracking any regulatory inquiries in the EU and US, as well as whether other watchdog groups will conduct independent audits. The episode underscores the growing need for transparent, verifiable safety standards before AI tools are marketed to under‑18 users.
Sources
Back to AIPULSEN