OpenAI to limit weapons information for teenagers in ChatGPT — Channel NewsAsia

OpenAI said it will tighten restrictions on weapons information that ChatGPT provides to teenagers. The new rules will apply to practical information about weapons and security systems that could be misused, even if a query shows no signs of violent intent.

Restrictions for teenagers

As Channel NewsAsia reports, the changes are part of OpenAI's new safeguards aimed at countering real-world violence. The company has also strengthened its user age-prediction system. Accounts of people whom the system considers to be under 18 will be moved to the service's teen mode.

OpenAI plans to notify parents if their child's account is temporarily suspended because of repeated conversations about violence. The company said the new safeguards will begin to be rolled out over the next few weeks and will continue to be refined.

More current news is available on the UA.News Telegram channel Telegram.

Measures for all users

For all users, ChatGPT will be able to automatically end conversations if a person, in the system's assessment, is obsessively focused on the topic of violence. After several such chats, OpenAI may temporarily suspend the account.

The company is also limiting details about past mass attacks, including information about the weapons used and ways of obtaining them. Previously, OpenAI said in its first report on teenagers' use of ChatGPT that they spend less than 15 minutes a day on average using the service.

At the same time, the nonprofit organization Common Sense Media stated in its report that the teen version of ChatGPT often notified parents late, or did not notify them at all, about conversations concerning suicide, self-harm, or eating disorders.