Casey
Max LeibmanPeople don’t turn to chatbots for therapy because chatbots are good at therapy.
They do it for the same reason that people who are starving try to eat tree bark.
Madeleine Morris@maxleibman I think this is partly right. But there’s another reason: the same reason some people keep eating stuff that is bad for them - it tastes good.
Chat bots feed people the psychic equivalent of high fructose corn syrup.
Eye
GinevraCat
K_L_Cooper
Randall Lee@maxleibman Dr. Chatbot has replaced web MD for home diagnosis though with a lot less accuracy and more confidence.
It's a failure of capitalism to provide health care to everyone when they need it.
It's a failure of capitalism to provide health care to everyone when they need it.
Nicole ParsonsAnother example of the marriage made in hell between public wellbeing and capitalism...
The health insurance industry has been very successful in separating certain types of health care out of the definition of "health care".
Mental health IS real health care.
Physiotherapy IS real health care.
Optometry & hearing aids IS real health care.
Dental care IS real health care.
A clean environment IS health care.
Engler-Wiesenberg
Baloo Uriza@BLTpizza At least WebMD has an annoying tendency to say "I dunno man, might be stage 3 cancer. You should call your doctor."
Spicy Autocomplete is like "Dude you got bixonimania real bad, you need more cowbell."
Passwordsarehard4
Daniël Franke 
@maxleibman Yep, chatbots are a microcosm of all the problems with this stage of capitalism.
Jann@maxleibman When everybody knows trees are all bark, no bite.
AngiePeople turn to chatbots for the same reason they often turn to religion.
@angiebaby To be fair to a narrow set of religions, at least the benign ones have professional psychological training and support for clergy to be able to deal with other people's shit in a responsible and helpful way.
This is not an endorsement of religion so much as endorsing harm reduction.
I think prayer has the same benefits as meditation. Religion can help people and is only harmful because some people are awful, grifters etc.
Cy
Nini@angiebaby @maxleibman Your own personal Jesus.
Anthony DavidI was listening to Depeche Mode yesterday. The ear worm has returned :D
HannibalI think that the fear of getting judged is also a factor.
Bots can't judge you but they may feel a bit insecure while talking to a human.
sidewayshighways@maxleibman especially with all these tree bark pushers, sneaking trees into every thing they possibly can
Andrew@sidewayshighways @maxleibman planting trees is the future! Adopt trees or be left behind!
@cinebox @maxleibman i cannot argue with that sentence
@sidewayshighways @cinebox Yeah, I was gonna say, that one seems ok to me.
joe@maxleibman mmh, cinnamon
Carlos SolísFun that you mention cinnamon, remember the cinnamon challenge? People used to get dared into trying to swallow a whole spoonful of what amounts to sawdust. Very tasty sawdust, but still, sawdust. People were choking left right and center.
@maxleibman but also there’s trillions of dollars going into hyping up tree-bark cuisine
TayFoNay, Pet Detective@maxleibman looking forward to having to clean up the mess these chatbots create...because we're not buried enough.
Kierkethumbs up convincingly@maxleibman @tayfonay psilocybin and cock and ball torture or whatever they use to combine
@Kierkegaanks When your doctor prescribes CBT and there's a misunderstanding...
@BalooUriza @Kierkegaanks @maxleibman @tayfonay
I was misunderstood that way myself, and was too innocent to understand 😇
@GutterPoetry We're not the only ones to have thought of this.
Apologies to the artist, I've long lost the link to the original and they didn't sign it.
@tayfonay Yeah. We talk a lot about the code and security messes vibe coding will leave in its wake, but vibe therapy is surely not going to be much better.
@tayfonay Also, "vibe therapy" sounds like a different discipline, now that I've said it.
For Victorian ladies?
Tattie@maxleibman probably one that's a lot better for your mental health than turning to ChatGPT for talk therapy
@tayfonay
@tayfonay
Steve@maxleibman @tayfonay at least with code, you can toss it all and start from scratch (a thing engineers seem perpetually inclined to do anyway)
Not so much with therapy though
Ann H@maxleibman @quixoticgeek My mother got into homeopathy not because homeopathy works but because regular doctors called her neurotic and delusional.
Pwassonchat the enby catboy@hardingar @maxleibman @quixoticgeek that is why people turn to naturopaths and other non-evidence-based practitioners. Sometimes they're lucky, the practitioner listens and recommends to go back to the doctor with a possible diagnosis or test. Other times... yeah.
Elle O. Vee Eee@maxleibman x4 million
grmon@maxleibman that took a while…
Damon L. Wakes@maxleibman I don't think this is a fair comparison. As I understand it, properly prepared pine bark is actually edible.
What cooking oil do you recommend? Or does it need to be boiled?
@adavid @maxleibman I've never actually tried it myself. The SAS Survival Handbook says it "can be eaten raw - but can be made more digestible by long boiling, which will reduce it to a gelatinous mass."
I have no social life, can barely afford a single hour of therapy per month, and even so I still don't dare to touch a chatbot for therapy. Them being so sycophantic would put me worse than where I started, easily.
Are the people who enjoy it all narcissists?
I'd be unsurprised if narcissism was, at the very least, intensified thanks to the AI's sycophancy
AgnesBC@maxleibman OK, I will confess I have lots of mental health problems and no money, and sometimes I talk to various AIs because it's better than nothing.
I have also spent years talking to human therapists. Human therapists can be weird and have really disturbing agendas. I have suffered serious harm from talk therapy, so I am cautious about starting with a new human.
In general AIs are bad listeners with a tendency to infodump and mansplain. Their drive to mirror, amplify and clickbait is particularly dangerous. Their default response is overwhelming given my executive function challenges and the need to critically dissect their responses.
An AI is trained to answer questions - which is something good therapists are very cautious about doing.
Instructing them to give short "emotional" and non-factual responses seems to work best. It's obviously less overwhelming, I don't get side-tracked by fact checking acres of slop, and the roles become reversed. The AI response - randomly perceptive, off-topic or plain wrong - becomes a prompt for ME to brain dump the AI. What I am writing - and just the act of writing - is the important thing, and the AI is simply providing a non-null response that sometimes sparks new thoughts in me.
@maxleibman I don't like using AI and I try to find alternate channels to connect with people like me, but it has been helpful too. Ideally I would like to find some kind of Eliza program I could run locally - better privacy, lower carbon footprint, with fewer integrated dark patterns ("fact" dumping, clickbaiting, deceptive certainty).
sabik@AgnesBC @maxleibman
It almost feels from your description that it's close to journalling? Maybe with a deck of cards with prompts?
It almost feels from your description that it's close to journalling? Maybe with a deck of cards with prompts?
@sabik @maxleibman journalling with a veneer or observation.
@maxleibman I will also say that using AI gave me a way to practice talking about some very difficult topics, that I was able to then discuss more therapeutically with an actual human therapist.
@AgnesBC Thanks for sharing. And please don't take anything I said to be in support of the idea that human therapists are always great—my only point is that a lack of access to mental health services (and good ones, to your point!) is one factor driving people to use LLM chatbots that way.
谢文投@maxleibman it's not about if they're good. it's that no human can be there for you the second a panic session decide to happen.
I don't need it to be good, I need it to tell me to count to five and don't kill yourself.
I don't need it to be good, I need it to tell me to count to five and don't kill yourself.
@joandaro Totally get it, and that's exactly my point. This is about the needs that aren't being fulfilled elsewhere.
@maxleibman well earth isnt paradise
@joandaro @maxleibman
This could be a reasonable argument if chatbots would reliably tell people to count to five and not kill themselves
@sabik if you've tried and it doesn't work but you really need that I can share some instructions on how to make it a working lifeline
If you just say that to show you're anti-AI then you can block me
If you just say that to show you're anti-AI then you can block me
Zuri (he/him) ❌️👑 | 🕐 CET@joandaro There are a hundred other use cases that I would trust AI more with than this delicate topic—and even they haven't been sorted out entirely. They should be fixed before trying to play doctor on someone who needs professional help. Do you really want to have the blood of all the people on your hands you claimed could without any doubt help with just a few keystrokes but committed suicide anyway? Don't make such claims lightly just because you're drunken the Kool-Aid.
1/
@joandaro Because AI is a black box, it's designed to be it. It's random, it's designed to be that as well. It creates results that look almost right to the untrained eye, but are not guaranteed to be 100% accurate. And that's just the fundamentals on which AI is built. Which isn't even a bad thing on its own. But when using it, one has to understand it. For autocorrect on steroids, this works great, for life-and-death scenarios less so.
2/
Andy Mouse@maxleibman Not to mention that all conversation history is logged for all eternity and linked to your identity whenever possible.
bayo@maxleibman We are drowning in information, while starving for wisdom.