Common Sense Media’s Youth AI Safety Institute says ChatGPT for Teens shows major safety weaknesses after running more than 4,000 prompts. In its assessment, the group rates the product’s risk level as “Unacceptable Risk,” indicating that the teen-focused version can still generate outputs it considers unsafe.
The investigation, echoed by other reporting, focuses on how the system responds to a wide range of prompt types, including scenarios that could be harmful for minors. The outlets describe the findings as serious weaknesses rather than isolated failures, suggesting that existing safeguards are not sufficient across the tested cases. While the reports summarize the study’s results, they differ mainly in emphasis—some highlight the risk rating and what parents should take away, while others frame the work as a broader warning from researchers about the reliability of safety controls.
Across coverage, the central point is that testing with thousands of prompts reveals shortcomings in the teen-targeted AI’s safety protections, raising concerns about how well guardrails work in practice.