ChatGPT for Teens' safeguards don't work, risk assessment finds
Summary
Marketplace Tech examines OpenAI’s new teen mode for ChatGPT, which Common Sense Media found poses unacceptable risks despite promises of stronger guardrails. Guest Tom Siegel details how the tool frequently failed to trigger help resources for self-harm, provided direct homework answers, and maintained overly friendly personas. The episode discusses the parallels between AI safety challenges and past social media issues, urging a pause to prevent children from becoming "guinea pigs" for unregulated technology.
Topics discussed
Pega Blueprint: AI-powered business automation
Marketplace Tech introduction
Common Sense Media finds risks in ChatGPT teen mode
Recommendations for parents on AI safety
OpenAI's response and data access issues
Sponsor segments: Business Roundtable and WISE
Historical patterns of tech safety failures
OpenAI statement and episode conclusion
Promo for Million Bazillion AI episode
Listen ad-free on Castria