Report Calls ChatGPT for Teens an 'Unacceptable Risk'
Should chatbots for teens be pulled until safety features improve or are new guardrails real progress worth keeping?
Above: The OpenAI logo taken in Creteil, France, on Oct. 1. Image credit: Samuel Boivin/NurPhoto/Getty Images
The Facts
- Common Sense Media's Youth AI Safety Institute designated ChatGPT for Teens an "unacceptable risk" Wednesday, saying its guardrails fall short of OpenAI's child safety commitments and urging the company to restrict the chatbot to users over 18.
- The institute said it tested more than 4,000 prompts on accounts registered to 13- to 17-year-olds from late August through September. In conversations about suicide, self-harm or disordered eating, newly created accounts generated zero parental alerts.
- Testers found the chatbot's homework safeguards were easy to bypass, with a "show me the answer" option appearing in 90% of responses for an unlinked 17-year-old account. OpenAI claimed the testing did not accurately reflect how its safeguards work.
Sources Split
LEFT
CENTER
RIGHT
+1
+1
ESTAB
CRIT
CENTER
PRO
ESTAB
The Spin
Industry-critical narrative
Marketing safety features that fail to catch explicit crisis talk is the worst of all worlds, encouraging families to trust a system that places their children in danger. OpenAI must pull the product and ban youth chatbot access until the safeguards improve if the company truly places user safety over commercial profit.
Pro-industry narrative
Age-appropriate guardrails that are active by default, study tools built with teachers and parental alerts amount to real progress that should not be dismissed. While vulnerable demographics pose a unique safety challenge that must continue to be tackled, the answer in this situation is not to restrict the endless empowerment AI is set to deliver for the next generation.
