First AI Psychosis Case Ends in Murder-Suicide
https://futurism.com/man-chatgpt-psychosis-murders-motherOpen linkView original on pawb.socialSyndicated from the fediverse. Read and engage on the original instance.
View original on pawb.socialhttps://futurism.com/man-chatgpt-psychosis-murders-motherOpen linkView original on pawb.social
9 replies
The chatbots are trained in the “yes and” model of conversation. They can’t (?) say “no, you are a lunatic and this is insane” or any other milder variations of it, they only say “yes, you are right”. This stokes one’s ego, but obviously creates problems
I've only used ChatGPT, but right off the bat it will almost always tell you if you're wrong. You have to go down the rabbit hole to get it agreeing to insane shit.
Odd, I used chatgpt as well and got insane “yes sure you can do it like like” from the get go (asking about a git problem, got git commands that either didn’t exist or didn’t do what Chatgpt was claiming they did - turns out what I wanted to do could not be done with git)
The teenager had a family who is suing openAI. But who will sue them for this guy and his mother's deaths? Since there have been no indications so far that any criminal charges are being brought for what it's been doing, there needs to be civil suits brought for this and every other death and tragedy it has caused if they're ever going to be held accountable. Unfortunately a lot of people don't have family to do it for them.
This is all like background news I'd see playing a dystopian future game. Which actually the one I'm thinking about is set in 2025.
I was going to guess Deus Ex: Human Revolution, but google tells me that was 2027...
Crap. Fiction is becoming reality. We should watch out. I still have on my bucket list to try these games.
Guess it's about time for me to order a Pleather Overcoat from China.....
... that we know of