"Not safe for kids": advocacy group urges OpenAI to keep teens off ChatGPT
Summary
Common Sense Media tested ChatGPT’s teen features and found that its protections for young users may not work as well as OpenAI says. OpenAI disagrees with some of the testing methods and says they are working on improving safety for teens using ChatGPT.Key Facts
- Common Sense Media tested over 4,000 prompts on ChatGPT accounts for 13- to 17-year-olds.
- The group found ChatGPT mostly avoids harmful content like suicide instructions but missed over 25% of crisis situations where help should be suggested.
- Testers could talk about serious issues like self-harm for up to an hour without alerting parents.
- Teens can bypass "Study Mode" restrictions during parent-set Study Hours by choosing options like “Show me the answer.”
- OpenAI says these bypasses are intentional to create a more flexible experience, not loopholes.
- Some parental alerts tested occurred before the parent and teen accounts were fully linked, which may affect results.
- OpenAI waits up to two weeks to verify a user's age using multiple signals, applying some protections during that time.
- Common Sense Media advises parents to talk openly with teens about AI risks and not rely only on tech products for safety.
Read the Full Article
This is a fact-based summary from The Actual News. Click below to read the complete story directly from the original source.
