It it not actually unpopular in the broader population, but AI seems like a dirty word on lemmy, just like a far-right’s gut reaction to words like “communism” or “socialism”.
It it not actually unpopular in the broader population, but AI seems like a dirty word on lemmy, just like a far-right’s gut reaction to words like “communism” or “socialism”.
Yes, thats why corporate control is bad. The problem is corporations, not the tech itself.
There is no substitute for professional help. AI for depression and lo eliness is like heroin for pain management - feels good, but isn’t doing you any favours and might just be the thing that ends you.
A Psychiatrist isn’t judging you, you are getting a diagnosis, therapies and treatment. It doean’t always feel as easy, because help isn’t always easy or comfortable.
Unfortunately “professional help” is not some deity that’s above human bias.
For the medicinal aspect, all they do is keep trying a bunch of med that do not work.
For the talking aspect, its nothing very useful, regardless of who it is. Is the “oh yeah that must be so hard for you”, either read from an instruction manual or reduce things down to “just go outside and take a walk” “just do x”. Meanwhile the front desk person tells you “hold on, you got a bunch of copays, go to the billing office” WHAT, when was that a thing? They never even said anything for the past several times. Pretty sure the insurance was supposed to cover that.
And frankly, your comment seems very Eurocentric. Like EVEN in a western democracy, healthcare (ahem 'Murica) isnt free. Disability Medicaid/Medicare (whatever the term is) is not easy to get. Its a bureaucratic nightmare.
And think about all the countries in um…
Africa
Middle East
Southeast Asia & South Asia
LOTS of places dont have the luxury of healthcare, and they have cultures that actively stigmitize mental health. Like I know my parents homeland, People’s Republic of China, has a horrible stigma of mental health. An open source AI model would be REALLY HELPFUL for those in these ultraconservative societies.
(Of course beta software in testing isnt stable, but if we have EFF or Wikimedia Foundation develop it (example organizations, not saying they would be the best to develop such a thing), eventually we could have a stable usable model)
It sounds like you have no idea WTF you’re talking about. This is the response I would expect from someone that has just heard from other people that have never gone to a therapist, what therapy is.
I’m certainly not a proponent of AI, but AI for therapy is arguably the absolute worst use case. Therapists and psychiatrists are medically trained to work in their professions. As far as drugs, yea many times it’s not the right drug it disagree on the first attempt because everyone is different and not everyone responds the same way to the same drug or dosage.
Depression is a condition that is crippling, and for some, anti depressants are the solution due to chemical imbalances. I personally know a handful of people that have gone through the ringer to find what works for them, and they are still alive today because of it. No AI can do that.
I do agree that the cost is an unfortunate hurdle, and I personally have been affected by that. Some non empathic and non human AI that will just add likely tell a person to kill themselves as give God advice is not a solution.
If you had cancer, would you turn to some non profit AI to give you a treatment, or would you go to an oncologist instead? It’s the exact same thing when dealing with mental issues.
I dont argue that AI is a cure, its more like a bridge to feel more stable and so they can be mentally sane enough to continue to seek self improvement, whether it be going outside, trying to socialize at a club, or employment. Because having an AI, feels like having “someone being with you” so you wouldnt feel anxious at the world. At that would be least if such an AI could be perfected, which of course currently these corporate backed AIs aren’t nearly as good as it could be, on top of the privacy issues.
This is the problem. AI isn’t that. There is no knowledge in an LLM. It’s impossible to train them in such a manner as how you see them being beneficial. Using it as a bridge is like getting blackout drunk to get to the next day. For the people that think they’re finding solutions, it’s more of an addiction.
This is therapy in a nutshell. Being able to do that effectively takes special training and situational awareness that LLMs can’t do. In your last post, specifically the part I quoted in my response, was you griefing therapy for exactly what your saying LLMs can do.
I encourage you to look into the real world anecdotes of people that have used AI for this purpose, or the people that have created significant others.
It would be great if it were possible for someone that’s struggling to ask a question of their computer and get all of the answers, but that’s nothing but wishful thinking. Even if the perfect model were trained with the perfect material, and somehow had no bias, it still wouldn’t be a bad choice, inherently because the technology just isn’t capable of doing what you’re looking for. It requires a human mind with empathy and true understanding to be able to guide someone, not a computer that is finding the closest word match based on a users input. There’s no intelligence or understanding in AI, just best guess to the next word.
There is no way to control chatbots enough to make them safe, even if they weren’t under capitalist control the technology just does not work that way. An emotionally vulnerable person, one who’s lonely or depressed, is exactly who you don’t want relying on a chatbot. Whether it spins into some ego trip or encourages suicide or steers them towards murder. Whatever it is there’s no thought there, it doesn’t understand it’s words. There’s zero way to keep it in check. The “guardrails” marketing materials claim are on chatbots now are just suggestions to the things.
That’s why the “my grandma” hack and such worked so well, because the thing has zero understanding. It doesn’t know it’s causing harm. Even one created and controlled by people who genuinely mean well it would not be a good idea.