OpenAI has expanded and consolidated its teen protections for ChatGPT, combining age prediction, default content limits, parental controls and model-level behavior rules for accounts believed to belong to users under 18. The July 16 update, highlighted by Crypto Briefing, is chiefly a status report on safeguards OpenAI has introduced since late 2025 rather than a single new product launch.
ChatGPT’s age-prediction system is the central change. Per OpenAI, it evaluates account signals including general conversation topics, usage timing, account activity and account age to decide whether an account may belong to a minor. If the system is uncertain, it can place the account in the more restrictive teen experience. Adults incorrectly classified as teens can complete age verification through Persona using a live selfie, government ID, or both, depending on location.
Accounts in the teen experience receive additional limits by default. OpenAI says these include reduced exposure to graphic violence, sexual or romantic roleplay, violent roleplay, risky viral challenges, extreme beauty content, unhealthy dieting material and body-shaming content.
The company’s December 2025 Model Spec update also added under-18 principles intended to shape how ChatGPT responds to teens. These rules call for more care around self-harm, suicide, dangerous activities and substances, sexualized interactions, eating disorders and requests to conceal unsafe behavior. OpenAI says the system should steer younger users toward real-world support rather than act as a substitute for trusted adults or professional care.
That distinction matters: these are behavioral and content safeguards, not a guarantee that every response will be suitable. OpenAI’s own help documentation notes that age prediction can make mistakes.
OpenAI updated those controls on July 13, adding a narrowly scoped notification when a linked teen account is banned for violent activity. The company says the alert is not meant to trigger on fictional writing, gaming, news or political discussion, general anger, or abstract questions. Parents can also enable Study Mode by default for new chats through the parental-controls settings.
OpenAI had previously said that, in rare instances of acute distress where a parent cannot be reached, it may involve emergency services or law enforcement. The company describes that as a last-resort safety measure.
Families using ChatGPT should link teen accounts through parental controls rather than rely solely on a birth-date entry, while adult users placed in teen mode can use OpenAI’s verification process to restore adult settings.
ChatGPT’s age-prediction system is the central change. Per OpenAI, it evaluates account signals including general conversation topics, usage timing, account activity and account age to decide whether an account may belong to a minor. If the system is uncertain, it can place the account in the more restrictive teen experience. Adults incorrectly classified as teens can complete age verification through Persona using a live selfie, government ID, or both, depending on location.
What the teen experience blocks
Accounts in the teen experience receive additional limits by default. OpenAI says these include reduced exposure to graphic violence, sexual or romantic roleplay, violent roleplay, risky viral challenges, extreme beauty content, unhealthy dieting material and body-shaming content.The company’s December 2025 Model Spec update also added under-18 principles intended to shape how ChatGPT responds to teens. These rules call for more care around self-harm, suicide, dangerous activities and substances, sexualized interactions, eating disorders and requests to conceal unsafe behavior. OpenAI says the system should steer younger users toward real-world support rather than act as a substitute for trusted adults or professional care.
That distinction matters: these are behavioral and content safeguards, not a guarantee that every response will be suitable. OpenAI’s own help documentation notes that age prediction can make mistakes.
Parental controls and new notifications
Parental controls, introduced in September 2025, let a parent or guardian link an account with a teen’s account through an email invitation. The linked adult can manage selected settings, including sensitive-content reduction, memory and chat-history options, and certain feature access. The controls do not provide access to the teen’s conversations.OpenAI updated those controls on July 13, adding a narrowly scoped notification when a linked teen account is banned for violent activity. The company says the alert is not meant to trigger on fictional writing, gaming, news or political discussion, general anger, or abstract questions. Parents can also enable Study Mode by default for new chats through the parental-controls settings.
OpenAI had previously said that, in rare instances of acute distress where a parent cannot be reached, it may involve emergency services or law enforcement. The company describes that as a last-resort safety measure.
Why Windows users should care
For Windows households, the practical impact is largely browser and app based: ChatGPT accounts classified as teen accounts may see different answers and unavailable interactions on the same PC that an adult account can use. There is no indication that these tools replace Windows Family Safety, Microsoft Edge controls, network filtering, or school and enterprise device-management policies.Families using ChatGPT should link teen accounts through parental controls rather than rely solely on a birth-date entry, while adult users placed in teen mode can use OpenAI’s verification process to restore adult settings.