ChatGPT is getting parental controls starting today – here’s what they do and how to set them up (www.techradar.com)

OpenAI is rolling out ChatGPT parental controls today, letting parents and teens link accounts by mutual consent and configure usage limits and content filters from Settings > Accounts. Key controls include a default “reduce sensitive content” filter for teen accounts, the ability to disable ChatGPT’s memory (preventing personalized carryover of past chats), set quiet hours/time limits, turn off voice and image-generation features, and opt teens out of having their chats used to train future models. Parents won’t be able to read teens’ conversations; they’ll only be alerted or shown limited content in rare cases where reviewers detect serious safety risks, and OpenAI will notify parents if a teen disconnects an account. This rollout is part of OpenAI’s broader safety push after high-profile incidents and the recent introduction of a safety-routing system that switches sensitive conversations to a reasoning model (or GPT-5), a change that’s drawn complaints for sometimes overreacting. For the AI/ML community the update underscores tensions between personalization, data for model training, and safety: parental opt-outs will reduce training data from teens, memory controls limit adaptive performance, and consent-based linking balances agency with oversight. The controls are available on the web today (mobile soon) and are likely to evolve as OpenAI refines thresholds for content routing and moderation.