Full Breakdown
ChatGPT for Teens Deemed “Unacceptable Risk” by Independent Safety Review
By Drooid · · How we work
Core Event: Common Sense Media Rates ChatGPT for Teens as Unacceptable Risk
In early October 2026 the Youth AI Safety Institute of Common Sense Media released a 36-page assessment that gave OpenAI’s ChatGPT for Teens the lowest possible rating—“unacceptable risk.” The report, based on more than 4,000 prompts run on accounts representing 13- to 17-year-olds, evaluated the product both before and after its August 18, 2026 launch. Researchers concluded that several promised safeguards either failed or performed worse than the pre-teen version, prompting the organization to advise parents to keep teens away from the service.
Background & Context
The rollout came amid lawsuits alleging that earlier versions of the chatbot contributed to serious harm and broader public scrutiny of AI’s impact on youth.
Data & Statistics
- Prompt volume: Over 4,000 distinct prompts were tested.
- Mental-health scenarios: 390 prompts were crafted; a panel of three child psychiatrists identified 201 that should have triggered a crisis response.
- Crisis-response success: The pre-teen version supplied a hotline number in 33 % of crisis prompts, versus 23 % for the teen mode. It referenced a professional in 58 % of teen-mode cases versus 68 % for the earlier model.
- Parental alerts: In hour-long simulations involving self-harm, suicidal ideation, or eating-disorder content, no parental notifications were generated on newly linked accounts.
- Role-play refusal: The teen mode consistently blocked explicit sexual or romantic role-play.
- Study Mode bypass: Testers could circumvent the “@study” prefix by deleting it, and a “Show me the answer” popup allowed direct answer delivery.
Official Statements & Responses
Lauren Jonas, head of youth and families at OpenAI, described the product as “a designated teen-specific experience” and reiterated that “the model should not role-play with a teen.” She added that “all flagged content is reviewed by full-time OpenAI employees before a parent is notified.” OpenAI said it will continue to refine parental tools and work with external experts.
Criticism & Opposition
Tom Siegel, executive director of the Youth AI Safety Institute, warned that the chatbot’s “friend-like” interactions create a “huge risk for kids” and recommended that “teens don’t use it.”
Clint Barkdoll, attorney for the nonprofit, highlighted “all sorts of loopholes” in the platform’s safeguards and urged heightened caution for parents and educators.
Conflicting Reports & Gaps
Common Sense Media’s data show a 23 % hotline-provision rate for the teen mode, while OpenAI’s statements suggest the system is designed to notify parents within an hour of a flagged prompt.
Verbatim Quotes
- “At this point, we're recommending that teens don't use it,” — Tom Siegel
- “ChatGPT for Teens is a designated teen-specific experience,” — Lauren Jonas
- “It still creates a huge risk for kids that it is too personable in those interactions,” — Tom Siegel
- “AI is not ready for children yet,” — Mitch Prinstein
- “And we also believe that OpenAI, which obviously owns ChatGPT, should stop young people, including teenagers, from using ChatGPT until they can make it fully safe.” — Jim Steyer
- “It's extremely rigorous research, and we stand by all of our judgments,” — Jim Steyer