Introduce Trusted Contact Safety Feature

## What changed On May 7, 2026 OpenAI announced Trusted Contact, an optional safety feature for adults using personal ChatGPT accounts. Users select a trusted contact in settings and send an invitation. Once accepted, automated systems and trained reviewers can notify that person if they detect serious safety concerns such as talk about suicide.

The feature rolls out gradually over coming weeks in supported regions. It stays unavailable for Business, Enterprise, or Edu workspaces.

## Why it matters OpenAI is placing a direct bet on layered human oversight for high-risk personal conversations. This move acknowledges that solo users sometimes turn to ChatGPT during vulnerable moments and attempts to route them toward real-world support.

It also creates a new privacy surface. A designated contact gains visibility into flagged activity, which may deter users who value complete confidentiality or push them toward temporary chats and deleted history.

## How to use it Open Settings inside a personal ChatGPT account, locate the Trusted Contact option, and enter the email of the person you want to invite. They must accept before any notifications can occur. The setting appears only for eligible personal accounts in supported regions.

No paid plan is required. Business or workspace accounts cannot access it.

## Watch for Confirm the bet if actual check-in rates rise and false-positive notifications stay low. The feature breaks if users treat it as surveillance and migrate to other models. Expect OpenAI to test similar contact flows for additional risk categories next.