"When he spoke to me about it, it was to tell me that he no longer saw any human solution to global warming,” his widow said. “He placed all his hopes in technology and artificial intelligence to get out of it”.
Suicide is a consequence of deep and prolonged psychological pain, in this case at least partially motivated by our frankly hopeless climate situation. This man had no faith in our collective ability to tackle the problem, and thus placed that faith in a magical tech solution as a form of coping. None of this is new, it’s basically the ideology of the portion of the ruling elite that’s not building bunkers in New Zealand, and runs downstream from there. What is new is the hype cycle for these LLMs. A lot of uninformed people think the singularity™ is around the corner. So it’s here, the thing that will fix the climate!
Now, everyone that’s tried chatGPT and it’s variants knows that it doesn’t really like to disagree with you. It was primarily trained to generate responses that please people. And the last thing a depressed person needs is to have their thoughts repeated by a third party unchallenged, like an automated form of rumination. I agree that additional barriers should be put in place to prevent the models from encouraging self-harm, but I also genuinely believe this guy using the actual 1960’s ELIZA would be less harmful. Because there’s no mysticism and hype surrounding it. What I’m trying to say is that OpenAI and every fucking “journalist” that covered ChatGPT uncritically have blood in their hands.
I feel like if I ever got too that ping, I’d hopefully want to bring a couple of the people dedicated to making things worse with me.
Spending six weeks asking my magic 8 ball if I should kill myself and keeping a tally of the responses
All A.I. shall be Nazified.
A lot of people in journalism and the general public, and even in the tech world where people ought to know better, think of AI as magic omniscience when it’s really just glorified autocorrect that will always give worse output than the material it was trained on. It’s a useful tool for automating routine work like writing emails, recognising patterns or debugging code but it isn’t real intelligence, let alone wisdom, and it should never be trusted blindly.
Whatever the social response to AI technology is, educating people about it being just another dumb tool is going to be part of it.
