ChatGPT will now contact police if it detects “imminent harm” from user

by

OpenAI is rolling out a number of updates to ChatGPT to make it safer for younger users, establish clearer rules about what the AI can discuss, and allow people to customize its behavior more easily.

These new guidelines are part of OpenAI’s broader effort to make its technology safer, more accessible, and more private for everyone, no matter their age.

Although the updates were largely aimed at teenage users, OpenAI explained that the new rules for the chatbot-regarding sensitive topics and personal data-apply to everyone, regardless of age.

Key changes coming to ChatGPT

A major change is that ChatGPT will now have stronger safety measures for users who are thought to be teenagers or younger.

This involves several steps, like preventing the AI from responding to flirty messages and teaching it to avoid discussing self-harm or suicide, even when these topics appear in stories or creative writing – especially when a young person is using the AI.

OpenAI uses a system to automatically restrict access to ChatGPT for users who appear to be underage. If a user’s age can’t be confirmed, the system assumes they are a teenager as a precaution.

The company clarified that human reviewers don’t read ChatGPT conversations unless there’s a significant safety issue, like threats of violence. If a young person appears to be struggling and their parents can’t be contacted, OpenAI will try to reach the parents and, if that fails and there’s an immediate risk of harm, will contact the authorities.

Although these limits currently only apply to users who are, or are believed to be, under 18, OpenAI explained that the goal is to create a safer experience for all users with clearer rules and increased monitoring.

OpenAI recently updated its policies following a lawsuit by Matt and Maria Raine. They allege that OpenAI’s software encouraged their son, Adam, to die by suicide.

Read More

2025-09-16 18:48