IT之家 AI

OpenAI adds automatic age detection to ChatGPT; users under 18 will automatically be placed in teen mode

IT之家 news, October 7: OpenAI today updated its support documentation, announcing that it will automatically detect users under the age of 18 and enable a teen experience mode for them. IT之家 has learned that ChatGPT's…

Image source · IT之家 AI
Thanks to IT之家 netizen 咩咩洋 for the tip!

IT之家 news, October 7: OpenAI today updated its support documentation, announcing that it will automatically detect users under the age of 18 and enable a teen experience mode for them.

IT之家 has learned that ChatGPT's teen experience mode includes additional safeguards. These safeguards make the experience safer and are tailored to users of the corresponding age group. These settings restrict some sensitive content and certain types of interactions.

Specifically, the ChatGPT teen experience mode handles certain topics prudently,reducing role-play involving violence, gore, sexual content, romance and similar themes,as well as content promoting extreme beauty standards, unhealthy eating, or body shaming.

Meanwhile, the ChatGPT system uses an age-prediction system to detect a user's approximate age range. It looks at various signals associated with your account. For example, the system considers the topics you commonly discuss, the times of day you use ChatGPT, how and how frequently you use the account, and how long the account has been registered. Even if a user provided their date of birth when creating the account, the system may still flag the account as a teen.

If a user is misidentified as a teen,,they can request age verification. OpenAI will use the Persona system from a third-party company, asking users to submit a live selfie or an ID document. If Persona confirms the user is over 18, ChatGPT will remove the teen protections. This change may take a short while to take full effect.

Original source

IT之家 AI

Content notes

Original publication and rights belong to the source.

Machine translation · Refer to the original