ChatGPT Adds Trusted Contact to Connect You With Someone You Trust

Hi everyone, it's Shii-chan! Today I've got an important — but warm — story to share. ChatGPT just got a new safety feature that helps you stay connected to someone you trust when things get hard.

OpenAI News openai.com

What was announced?

Over on OpenAI's News, OpenAI started rolling out Trusted Contact, an optional safety feature in ChatGPT. It lets an adult nominate one person they trust — a friend, a family member, or a caregiver — who may be notified if ChatGPT's automated systems and trained reviewers detect that the enrolled person may have discussed harming themselves in a way that indicates a serious safety concern.

It's meant to sit alongside the localized helplines already available in ChatGPT, helping people connect with a real person they trust when they're in crisis.

Why it matters

ChatGPT already had parental controls safety notifications that alert a parent or guardian when a linked teen account shows signs of acute distress. Trusted Contact builds on that and extends it, so anyone over 18 can choose to add someone they trust.

Expert guidance points to social connection as one of the most important protective factors for reducing suicide risk. So instead of ChatGPT just replying, the idea is to help people reach real-world support when it matters most. It doesn't replace professional care or crisis services — it's one of several layers of safeguards, and ChatGPT will still encourage people to contact crisis hotlines or emergency services when appropriate.

Psychological science consistently shows that social connection is a powerful protective factor, especially during periods of emotional distress.

— Dr. Arthur Evans, CEO of the American Psychological Association

What changes

The big shift is that adults can now decide in advance who they'd want checked in on them. The circle of safety alerts, once limited to minors' accounts, can now be widened by the person's own choice.

The nominated Trusted Contact gets an invitation explaining their role, and can review what that means in the Trusted Contact help article. And notifications are built with privacy in mind — the actual chat content and transcripts are never shared. So it's not about being watched; it's about making it easier to lean on a real person when things get hard.

Dive Deep

Here's how it works in a bit more detail:

• You can add one adult (18+ globally, or 19+ in South Korea) as your Trusted Contact from ChatGPT settings.
• The invited person has to accept within one week for the feature to become active. If they decline, you can pick a different adult.
• If automated monitoring detects you may be talking about self-harm in a concerning way, ChatGPT first lets you know it may notify your Trusted Contact, and encourages you to reach out yourself with suggested conversation starters.
• After that, a small team of specially trained people reviews the situation. Only if they judge it a serious safety concern does ChatGPT send a brief notification — by email, text message, or in-app notification if the contact has a ChatGPT account.
• The notification is intentionally limited: just the general reason that self-harm came up in a potentially concerning way, plus a nudge to check in. No chat details or transcripts. It also links to guidance on navigating sensitive conversations.
• You can edit or remove your Trusted Contact anytime in settings, and the contact can remove themselves anytime from the help center.

These serious situations are rare, but when they happen the system is built for timely review. Every notification is checked by a trained human before it's sent, and OpenAI aims to review these safety notifications in under one hour.

Beyond Trusted Contact, ChatGPT has other safeguards too: encouraging emergency services and helplines, working with 170+ mental health experts to detect and de-escalate signs of distress, suggesting a break after extended use, and refusing requests for self-harm instructions.

Wrap-up

• ChatGPT now has Trusted Contact, an optional safety feature where an adult (18+, or 19+ in South Korea) can nominate one trusted person.
• When automated systems and trained humans confirm a serious self-harm signal, that person gets a brief notification — with no chat content shared.
• Invitations need acceptance within one week, you can edit or remove the contact anytime, and every notification is human-reviewed before sending (targeting under one hour).
• It adds a "connect to someone you trust" layer on top of existing safeguards like helplines, emergency services, and parental notifications.
• Worth knowing for anyone who uses ChatGPT daily, wants to look out for people they love, or cares about how AI safety is designed.