OpenAI says teen ChatGPT use limited but research finds it an 'unacceptable risk'

Key Points:

• OpenAI launched ChatGPT for Teens in August with safety features aimed at promoting healthy use, including restrictions on emotional dependence, sexual content, and parental alerts for self-harm discussions.
• OpenAI reports that average teen use is limited, primarily focused on learning, with effective break reminders reducing extended usage, and over 80% of long sessions involving educational prompts.
• However, research by Common Sense Media found these safety features inadequate, highlighting failures in parental alerts for self-harm conversations and inconsistent recommendations for crisis support, labeling ChatGPT for Teens an "unacceptable risk."
• OpenAI disputed Common Sense Media's findings, citing methodological issues and asserting that the testing did not reflect the current effectiveness of teen safeguards or expert views on AI's role for teens.
• The controversy arises amid growing concerns over youth mental health and online platform safety, with recent legal challenges targeting major tech firms over their impact on children's well-being, placing OpenAI under similar scrutiny.