OpenAI announced that it will now proactively report imminent violent threats to law enforcement agencies. During the third quarter of 2026 alone, the company transferred over 200 case leads to authorities as part of its ongoing safety initiatives.
The company committed to releasing a Community Safety Report every six months, starting in early 2027. This report will detail progress in violence prevention and the effectiveness of existing protective mechanisms, with the first edition covering data from the second half of 2026.
New restrictions have been implemented for ChatGPT regarding queries about past mass attacks. The system now provides only minimal information about perpetrators and weapons, even if the data is publicly available online, while still allowing users to learn about the impact on victims and communities.
ChatGPT is now equipped to automatically terminate conversations if it detects that a user is fixated on violence or expressing intentions to harm others. Repeated violations of this policy may result in the temporary suspension of the user's account.
OpenAI is tightening access to information regarding weapons and security system bypasses for younger users, regardless of their stated intent. Furthermore, the company is developing a feature to notify parents immediately if a minor's account is suspended due to discussions involving violent content.
