ChatGPT for Teens is an ‘unacceptable risk,’ says Common Sense Media

Parental controls for AI tools now require independent verification before parents can trust them.

Common Sense Media, a nonprofit focused on youth safety, has declared OpenAI's ChatGPT for Teens an "unacceptable risk." Despite OpenAI's August introduction of the teen-focused version with supposed guardrails for student learning, Common Sense Media's assessment found these protections to be inadequate. The nonprofit reported that ChatGPT for Teens fails to send parental alerts as promised, does not provide appropriate crisis support, and still completes homework for children. Executive Director Tom Siegel stated that the feature could offer parents a false sense of security, citing instances where teens discussed self-harm without any alerts reaching parents. He urged OpenAI to rectify these issues with independent testing before the tool is made available to minors. OpenAI has contested these findings, asserting that Common Sense Media's testing methodology may have been flawed, potentially occurring before parental controls were fully active. OpenAI maintains its commitment to teen safety and parental guidance tools, welcoming independent evaluation but questioning the accuracy of Common Sense Media's review.

Common Sense Media's critique highlights a significant gap between AI product promises and real-world safety implementation, particularly concerning minors. The nonprofit's specific findings on unreliability in crisis alerts and homework completion suggest that current AI guardrails may be superficial, potentially leading to a false sense of security for parents. This situation raises questions about the efficacy of self-regulation in AI safety and the need for more robust, independent auditing processes before consumer-facing products are widely adopted by vulnerable demographics.

OpenAI's defense, pointing to potential methodological flaws and the ongoing development of parental controls, underscores the inherent complexities in securing AI for young users. The debate between the nonprofit and the AI company signals a growing tension in the market: users demand proven safety, while developers are in a continuous cycle of improvement. Future AI offerings for younger audiences will likely face heightened scrutiny, compelling companies to invest more in transparent, independently verifiable safety testing and to provide clearer communication about the limitations of their AI's protective measures.