OpenAI Restricts Weapons Information for ChatGPT Teen Users
OpenAI is implementing new safeguards to block weapons-related information for teenagers and automatically end conversations that fixate on violence.
OpenAI announced new safeguards for ChatGPT on Friday intended to prevent real-world violence by restricting access to weapons-related information for teenage users. These limits block practical details on weapons and security systems, regardless of whether a user's query appears harmless or lacks violent intent.
For all users, the chatbot will now automatically terminate conversations if a person appears to be fixating on violence. The company may temporarily pause accounts after multiple such occurrences. OpenAI is also limiting the amount of detail the AI provides regarding past mass attacks, specifically concerning the weapons used and how they were procured.
To enforce these measures, the company strengthened its age prediction system to ensure users under 18 are placed in the teen experience. OpenAI plans to notify parents if a teenager's account is paused. These updates will roll out over the next few weeks. The move follows reports from Common Sense Media that the teen version of ChatGPT often failed to alert parents in a timely manner regarding conversations about eating disorders or self-harm.
