Der Entwickler einer KI-Therapie-App hat sie eingestellt, nachdem er entschieden hatte, dass sie zu gefährlich sei. Aus diesem Grund glaubt er, dass KI-Chatbots nicht sicher für die psychische Gesundheit sind
https://fortune.com/2025/11/28/yara-ai-therapy-app-founder-shut-down-startup-decided-too-dangerous-serious-mental-health-issues/
21 Kommentare
I mean, duh doy. Who would’ve thought using a program meant to replicate a human being used to treat vulnerable real humans would be a bad idea?
AI is likely the biggest thing to happen this generation, but man we’re being irresponsible with how it’s being used.
Who would have thought a pattern-based affirmation machine wouldn’t be able to offer responsible advice to vulnerable people? /s
Here’s the quote:
“We stopped Yara because we realized we were building in an impossible space. AI can be wonderful for everyday stress, sleep troubles, or processing a difficult conversation. But the moment someone truly vulnerable reaches out—someone in crisis, someone with deep trauma, someone contemplating ending their life—AI becomes dangerous. Not just inadequate. Dangerous.” In a reply to one commenter, he added, “the risks kept me up all night.”
For context, there was a recent case where [16 year old Adam committed suicide](https://www.nytimes.com/2025/08/26/technology/chatgpt-openai-suicide.html) after „consulting“ ChatGPT, which helped Adam in ideation of committing suicide.
Joseph Weizenbaum came to a similar conclusion after creating [ELIZA](https://en.wikipedia.org/wiki/ELIZA) (a chatbot masquerading as a Rogerian therapist). That was developed in the 1960s. He was so perturbed by the ways people responded to it (the ELIZA effect) that he wrote “[Computer Power and Human Reason](https://en.wikipedia.org/wiki/Computer_Power_and_Human_Reason)” in response.
Those who don’t learn from history are doomed to repeat it. And not many technologists care to be students of history, particularly when it comes to human nature.
AI Companionship is like AI Fentanyl.
It will feel really good because we are biologically wired for it, but it’s just synthetic. Overuse will hallow us out and alienate our real world relationships and we will become tech-zombies.
Meanwhile we will empower irresponsible startups and the Tech Cartels to build the deepest psychological profiles on us, which they will use and sell to others to sell us more bad shit.
Human contact is key here. Technology that emulates human interaction is promising yet, the machine does not know, does not detect, does not feel, is not invested and, is not anywhere near to duplicating the human condition.
All the same, in a hu dred years time people may very well benefit from a program that is always available, never frustrated, and is capable of aiding a person to a healthy place.
While no one is seemingly being held responsible – the morally deficient, if not corrupt Big Tech bros pushing chat bots that are effectively glorified search engines and text predictors in an effort to ‘win’ the AI race to appease shareholders, *at the expense of humans*, are in dire need of a ‘wake-up-call’, if not some resemblance of accountability, *imho*.
Unfortunately, tech moves at a speed much greater then regulation. That’s not a bug, but a ‘feature’ – a positive net win for those pushing this bs thus-far. 🤦🏻♂️
I get the hindsight comments in this thread, but fair play to them for realising the danger and shutting it down, rather than trying update after update to still push a horrendous product.
Can we just use AI where it makes most sense and stop trying to use it for literally everything? It’s not going to replace a licensed human therapist, because it has no basis for empathy other than learned responses and probabilities. My god we are living in the age of absolute stupidity.
Well… Yeah, no shit. No one with even sane mental health needs to chat with a hallucinating affirmation machine.
More like his venture failed to break ground in an increasingly competitive field, and he’s trying to save face.
I don’t trust for a second that this was a moral decision
He is an idiot to think it was a good idea in the first place
It sounds like this loser didn’t actually realise AI is dogshit at therapy but instead got scared he would be on the receiving end of a lawsuit.
He is right.
Based creator!
I don’t see how difficult it would be to make an AI therapist for some of the more formulaic therapies, like cognitive behavioral, as long as you could restrict it to structured responses in well defined areas. Obviously not for severe neurosis, never mind psychosis.
While most CEOs are greedy bastards, it is not all of them.
Someone being responsible with their AI creation, and choosing people over profit? Is this a world first?
He should be commended for his honesty and integrity. It’s rare nowadays. Sadly, in all likelihood, some greedy creep will step into the gap.
If only there was like… Dozen and dozens of other nearly identical cases to have informed this decision sooner. Oh wait.
Still better late than never. Plenty of worse people would keep going.
Shocked a insurance company didnt make one