OpenAI to restrict weapons information for teens on ChatGPT
The company may also end conversations and suspend accounts over fixation on violence, and will strengthen alerts to parents.
Translated from the Portuguese original by the Panorama newsroom. Read in Portuguese
Summary written by the Panorama newsroom from the sources listed below. Facts are attributed to the original publications.
OpenAI announced on Friday (30) new safety rules for ChatGPT, focused on preventing real-world violence. According to g1, the company will further limit the weapons information provided to teenagers, even when questions seem harmless.
The chatbot will be able to automatically end conversations when it detects signs of fixation on violence. If this recurs, the user's account may be temporarily suspended. For teen accounts linked to their parents', OpenAI intends to notify the guardians when there is a suspension for this reason.
The company also said it has improved the system that estimates users' ages, so as to direct those identified as teenagers to the version aimed at under-18s. In that version, practical information about weapons and security systems that could be misused is more restricted.
For all users, ChatGPT will provide fewer details about past mass attacks, including the weapons used and how they were obtained. The measures begin rolling out in the coming weeks and will continue to be refined, according to the company.
g1 recalls that the announcement comes amid the debate over the safety of teenagers who use AI. This month, OpenAI said this group spends less than 15 minutes a day on ChatGPT, and Common Sense Media pointed to flaws in parental alerts.