ChatGPT Trusted Contacts is an optional safety feature that notifies a designated trusted adult if the AI detects discussions indicating serious self-harm or suicide risk. OpenAI introduced the feature in early May 2026 as a response to growing concerns about AI misuse linked to tragic outcomes, including overdoses, violence, and suicides that have sparked lawsuits against the company. Unlike automated crisis alerts, this system combines AI detection with human review before any notification reaches your contact, adding a deliberate layer of caution.
Key Takeaways
- ChatGPT Trusted Contacts is a free, opt-in feature available only to users over 18
- You can designate only one trusted adult contact who must accept an invitation within one week
- Notifications are limited to general context and do not include chat transcripts or direct quotes
- Human reviewers assess flagged conversations before any notification is sent
- Feature is rolling out gradually; check Settings to see if it is available on your account
How ChatGPT Trusted Contacts Actually Works
The system operates in four distinct steps. First, automated detection identifies conversations that discuss self-harm or suicide in a way that signals serious risk. Second, ChatGPT informs you that your trusted contact may be notified and encourages you to reach out for help directly, offering conversation starters. Third, a small team of specially trained human reviewers assesses the conversation—typically within one hour—to determine if the concern is genuine. If reviewers do not flag it as a serious safety issue, no notification goes out. Only if confirmed as a legitimate crisis does the system proceed to step four: sending a limited notification to your trusted contact via email, text, SMS, or in-app alert.
The notification itself is deliberately sparse. It shares only the general reason that self-harm came up in a potentially concerning way and encourages your contact to check in on you. Critically, it does not include chat details, transcripts, screenshots, or direct quotes, protecting your privacy even in a crisis moment. This design reflects OpenAI’s effort to balance intervention with confidentiality—your trusted contact knows you may need support, but they do not get access to the full conversation.
Setting Up ChatGPT Trusted Contacts
The setup process differs slightly between desktop and mobile. On desktop, click your profile name in the bottom-left corner, select Settings, and navigate to the Trusted Contacts menu to add your designated person. On mobile, tap your profile name, scroll to App Settings, and select Trusted Contact. You can only designate one trusted adult contact at a time.
Once you add someone, they receive an invitation via email, text, SMS, or in-app notification and must accept within one week. You retain full control: you can change, edit, or disable your trusted contact at any time. This flexibility means the feature remains truly optional—if you decide it is not right for you, disabling it takes seconds.
Who Can Use ChatGPT Trusted Contacts and When
The feature is available only to adult users over 18. OpenAI developed it with guidance from clinicians and safety experts, mimicking the parental safety alert system already in place for teen accounts. The rollout began in early May 2026, but availability is expanding gradually over weeks. If you do not see Trusted Contacts in your Settings or within conversations, your account may not have access yet—check back periodically as OpenAI continues to roll it out globally.
The feature complements existing localized helplines already integrated into ChatGPT. Rather than replacing crisis resources, it adds a personal layer: someone who knows you and cares about you gets a heads-up that you may need immediate support. For users in regions with robust mental health infrastructure, this creates a bridge between AI detection and human intervention.
Why This Matters Right Now
ChatGPT Trusted Contacts arrives at a moment of intense scrutiny over AI safety. Reports have linked AI interactions to tragic outcomes, prompting lawsuits and regulatory pressure on OpenAI. By introducing human review before any notification, OpenAI is attempting to reduce false alarms while still catching genuine crises. This is not a perfect system—no safety tool is—but it represents a deliberate attempt to add friction and human judgment to an otherwise automated process.
The privacy-first design is particularly notable. OpenAI could have sent full chat transcripts to your trusted contact, but chose not to. This decision acknowledges that privacy matters even in emergencies, and that your contact does not need every detail to understand you are struggling.
Is ChatGPT Trusted Contacts right for you?
If you live alone, struggle with mental health, or have a trusted person in your life who you would want to know if you were in crisis, this feature is worth enabling. If you value privacy above all else or do not have a trusted contact, leaving it disabled is also reasonable. The feature is opt-in precisely because OpenAI recognizes this is a personal decision.
How long does it take for a trusted contact to be notified?
The entire process—from AI detection through human review to notification—typically takes within one hour. This is not real-time, but it is fast enough to matter in a crisis. The human review step intentionally adds delay to prevent false alarms from reaching your contact.
Can my trusted contact see my full chat history?
No. Your trusted contact receives only a limited notification that explains the general reason self-harm came up in a concerning way. They do not get access to transcripts, quotes, or chat history. This protects your privacy while still alerting them that you may need support.
ChatGPT Trusted Contacts is a modest but meaningful step toward responsible AI safety. It does not solve the broader question of how AI should handle mental health crises, but it does add a layer of human oversight and personal connection that purely automated systems lack. If you are over 18 and have someone you trust with this responsibility, enabling it costs nothing and could matter in a moment when you need it most.
Where to Buy
Apple iPhone 17e | Apple iPhone 17e
Edited by the All Things Geek team.
Source: Tom's Guide
