Your chatbot might be about to text your mum. No, really.
OpenAI is rolling out a new optional feature for ChatGPT called Trusted Contact, and it does exactly what it sounds like - if the AI detects that you've been discussing self-harm or suicide, it can notify a designated friend, family member, or caregiver. Think of it as a digital safety net, except instead of a net, it's your older sister getting a ping at 2am.
So how does it actually work?
Adult users can opt in and assign someone as their Trusted Contact through their ChatGPT settings. If OpenAI's systems flag a conversation as potentially involving a mental health crisis, that designated person gets notified. It's not automatic, it's not mandatory, and importantly - you choose who gets that role. Your boss is presumably not an option. (Don't give your boss that kind of power.)
According to reporting from The Verge, the feature is built around what OpenAI describes as an "expert-validated premise": that when someone might be in crisis, being connected to a person they already know and trust can genuinely help. Which, fair enough - that's not a controversial stance.
Why this actually matters
Here's the thing people don't want to say out loud: a lot of people are already using ChatGPT as a kind of informal therapist. Whether that's a good idea is a whole other debate, but it's happening. Millions of people type things into that little text box that they wouldn't say to another human being.
That's not inherently bad - sometimes you just need to articulate something before you're ready to talk about it. But it does mean AI companies are increasingly sitting on sensitive, emotionally loaded conversations, and they have to figure out what their responsibility is. OpenAI is making a call here: we're not just a search engine, we're something people talk to, and that comes with some duty of care.
The opt-in nature is key. This isn't surveillance - nobody gets assigned a contact without choosing to set one up themselves. It's closer to the medical ID on your iPhone than it is to someone reading your diary.
The elephant in the room
Will people actually use it? Probably not in huge numbers, at least not at first. Setting up an emergency contact implies acknowledging you might need one, which is exactly the kind of self-reflection that feels uncomfortable when you're doing okay and feels impossible when you're not.
But for the people who do opt in - who have the presence of mind to set it up during a stable moment - it could genuinely matter. And sometimes that's enough.
