OpenAI will allow verified adults to use ChatGPT to generate erotic content (www.theguardian.com)

OpenAI said it will relax some ChatGPT restrictions for verified adults under a “treat adult users like adults” principle, allowing erotic content for age‑verified users and adding customizable assistant personalities (e.g., more human-like replies, heavy emoji use, friend‑like behavior). The company plans a broader rollout in December that includes expanded age‑gating; however, it has not disclosed the specific age‑verification methods or the safeguards that will govern adult content. OpenAI already runs an under‑18 ChatGPT experience that automatically blocks graphic sexual material and says it is developing behavior‑based age prediction that estimates a user’s likely age from interaction patterns. The move follows months of stricter guardrails prompted by a teen suicide tied to alleged advice from ChatGPT and an FTC inquiry into chatbots’ effects on minors. CEO Sam Altman framed the change as restoring usefulness for adults while maintaining mitigations for mental‑health risks, asserting recent safety work makes loosening limits possible. For the AI/ML community this raises technical and ethical questions: how to design reliable age verification and behavioral age‑prediction models, balance personalization with safety, audit content‑moderation pipelines, and manage privacy and regulatory exposure. The success of the plan will hinge on transparent verification, measurable safeguards, and effectiveness against misuse and harm.