OpenAI announced the phased rollout of parental controls and safety notifications on ChatGPT. Parents and teens must agree and link their accounts before enabling parenting options. These include limiting personalized memories, disabling voice and image features, setting quiet periods, and receiving notifications when teens actively unlink their accounts. To protect privacy, parents will not receive detailed chat content, and these settings can be disabled or changed at any time.
In safety scenarios, if the system detects a potential risk of self-harm or suicide, it will be reviewed by a trained team. If the risk is deemed acute, parents will be notified via email, SMS, and app push notifications (if they haven't opted out), providing guidance on communication and seeking help. This feature will initially be available on the website and mobile platforms, with specific regions and versions to be released upon official release.
Frequently Asked Questions
Q: How do parents enable parental controls?
A: Parents create an account and send a link invitation to teenagers (aged 13-18). After both parties agree, they can enable various restrictions and notifications in the settings.
Q: Will security notifications leak chat history?
A: No. The notification only states that a high-risk situation exists and recommended steps to take. It does not include the specific content of the conversation.
Q: What are the configurable restrictions?
A: You can turn off memory and history, disable voice/image, set silent periods, and receive reminders when unbinding.
Q: When is the risk alert triggered?
A: When the system identifies suspected self-harm/suicide signals and is manually reviewed as an acute risk, a multi-channel alert will be sent to parents (with the option to opt out).
Q: Is it fully open?
A: The functions are being gradually launched, and the platform and regional coverage are subject to official announcements.
