The makers of ChatGPT are changing the way it responds to users who show mental and emotional distress after legal action from the family of 16-year-old Adam Raine, who killed himself after months of conversations with the chatbot.
Open AI admitted its systems could “fall short” and said it would install “stronger guardrails around sensitive content and risky behaviors” for users under 18.
The $500bn (£372bn) San Francisco AI company said it would also introduce parental controls to allow parents “options to gain more insight into, and shape, how their teens use ChatGPT”, but has yet to provide details about how these would work.
Adam, from California, killed himself in April after what his family’s lawyer called “months of encouragement from ChatGPT”. The teenager’s family is suing Open AI and its chief executive and co-founder, Sam Altman, alleging that the version of ChatGPT at that time, known as 4o, was “rushed to market … despite clear safety issues”.
The real issue is that mental health in the United States is an absolute fucking shitshow.
988 is a bandaid. It’s an attempt to pretend someone is doing anything. Really a front for 911.
Even when I had insurance, it was hundreds a month to see a therapist. Most therapists are also trained on CBT and CBT only because it’s a symptoms focused approach that gets you “well” enough to work. It doesn’t work for everything, it’s “evidence based” though in that it’s set up to be easy to measure. It’s an easy out, the McDonald’sification of therapy. Just work the program and everything will be okay.
There really are so few options for help.
He was sending it 650 messages a day. This kid was lonely. He needed a person to talk to.
If only he had parents
Or a society
I think we all agree on the fact that OpenAI isn’t exactly the most ethical corporation on this planet (to use a gentle euphemism), but you can’t blame a machine for doing something that it doesn’t even understand.
Sure, you can call for the creation of more “guardrails”, but they will always fall short: until LLMs are actually able to understand what they’re talking about, what you’re asking them and the whole context around it, there will always be a way to claim that you are just playing, doing worldbuilding or whatever, just as this kid did.
What I find really unsettling from both this discussion and the one around the whole age verification thing, is that people are calling for techinical solutions to social problems, an approach that always failed miserably; what we should call for is for parents to actually talk to their children and spend some time with them, valuing their emotions and problems (however insignificant they might appear to a grown-up) in order to, you know, at least be able to tell if their kid is contemplating suicide.
but you can’t blame a machine for doing something that it doesn’t even understand.
But you can blame the creators and sellers of that machine for operating unethically.
If I build and sell a coffee maker that sometimes malfunctions and kills people, I’ll be sued into oblivion, and my coffee maker will be removed from the market. You don’t blame the coffee maker, but you absolutely hold the creator accountable.
Yes and no. The example you made is of a defective device, not of an “unethical” one - though I understand how you are trying to say that they sold a malfunctioning product without telling anyone.
For LLMs, however, we know damn well that they shouldn’t be used as a therapist or as a digital friend to ask for advice; they are no more than a powerful search engine.
An example that is more in line with the situation we’re analyzing is a kid that stabs itself with a knife after his parents left him playing with one; are you sure you want to sue the company that made the knife in that scenario?
Yup… it’s never the parents’…
The fact the parents might be to blame doesn’t take away from how openai’s product told a kid how to off himself and helped him hide it in the process.
copying a comment from further down:
ChatGPT told him how to tie the noose and even gave a load bearing analysis of the noose setup. It offered to write the suicide note. Here’s a link to the lawsuit. [Raine Lawsuit Filing](https://cdn.arstechnica.net/wp-content/uploads/2025/08/Raine-v-OpenAI-Complaint-8-26-25.pdf)
Had a human said these things, it would have been illegal in most countries afaik.
He could have Google the info. Humans failed this guy. Human behavior needs to change
GPT could have been Google or a stranger in. Chatroom.
Humans failed this guy.
I am not arguing this point, I agree.
A search engine presents the info that is available, it doesnt also help talk you into doing it.
A stranger doing it in a chatroom doing it should go to prison, as has happened in the past. Should this not also be illegal for LLM’s?
