OpenAI restricts weapons-related information for teens on ChatGPT
What's the story
OpenAI has announced stricter limits on the "weapons-related information" that teens can access through ChatGPT. The company already has policies in place to prevent the AI from assisting users in committing violent acts or teens acquiring weapons. However, as a further measure, OpenAI plans to "restrict practical information about weapons and security systems that could be misused to facilitate violence." This will be done even if there is no apparent intention of violence from its teen users.
Policy expansion
New restrictions will also apply to adult users
The new restrictions from OpenAI aren't just limited to teens. The company is also strengthening policy boundaries for all its users.
This includes queries about "reconstructions of past mass attacks, including details about the weapons and equipment used and how they were obtained, and detailed comparisons on scale of harm that could amplify an attacker's notoriety."
Safety measures
New safeguards build on existing parental controls
The new restrictions will work with the parental controls OpenAI introduced last year.
These controls notify parents if their linked teen accounts violate ChatGPT policies related to violent threats or acts of violence.
The additional safeguards come after a wrongful death lawsuit against OpenAI last year and its CEO Sam Altman's public apology for not alerting police about alarming chat logs related to the Tumbler Ridge shooting earlier this year.
Future plans
Other updates coming to ChatGPT
OpenAI plans to release its first Community Safety Report in early 2027, with updates every six months.
The company has also improved its age prediction system to move users it estimates are under 18 into the teen experience.
It will also notify parents if their teen's account is paused for repeatedly discussing violence.
These new safeguards will start rolling out in the coming weeks, and OpenAI intends to continue refining them over time.