OpenAI’s Teen Safety Features Will Walk a Thin Line
Summary
OpenAI has announced a set of teen safety measures for ChatGPT centred on an age‑prediction system that aims to detect users under 18 and route them to an “age‑appropriate” model that blocks graphic sexual content. The company will also roll out parental controls by the end of September so parents can link to their child’s account, set usage limits, manage conversations and receive alerts if the system detects acute distress. In cases of suspected self‑harm the system may notify parents and, if parents are unreachable and danger is imminent, contact authorities.
CEO Sam Altman framed the changes as an attempt to balance freedom, privacy and teen safety, but the announcement arrives amid increased scrutiny: the FTC has asked major AI firms about impacts on children and OpenAI remains under a court order to preserve user chats. Observers note that limited research exists on how teens actually use chatbots, leaving important questions about effectiveness and unintended consequences.
Key Points
- OpenAI is developing an age‑prediction system to identify under‑18 users and route them to a model that blocks graphic sexual content.
- Parental controls will allow parents to link accounts, set time limits, manage conversations and receive notifications about acute distress.
- If the system detects suicidal ideation or imminent danger it will alert parents and may contact authorities if parents cannot be reached.
- The move underscores a trade‑off between privacy and safety: OpenAI says it will prioritise teen safety while continuing to defend adult privacy.
- The rollout occurs amid regulatory pressure (including an FTC inquiry) and legal constraints, and critics say more research is needed on teen interactions with AI.
Why should I read this?
Because if you care about kids, privacy or where AI is heading, this one’s for you. OpenAI’s tweaks could change who sees what, who gets alerted when a teen is in crisis, and how much control parents and companies have over conversations. It’s a neat mix of sensible safety steps and awkward trade‑offs — quick read, big consequences.
Source
Source: https://www.wired.com/story/openai-launches-teen-safety-features/
