Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

OpenAI says over a million people talk to ChatGPT about suicide weekly

techcrunch.com

Post reply on HN
Print view View original

411–420 of 585 posts

Re: OpenAI says over a million people talk to ChatGPT about suicide weekly

#411

Post by ml-anon »

josephg wrote: post #341
alex-moon wrote: post #317

Earlier quoted context omitted.

I suspect you've never done therapy yourself. Most people who have worked with a professional therapist understand intuitively why the only helpful feedback from an LLM to someone who needs professional help is: get professional help. AIs are really good at doing something to about 80%. When the stakes are life or death, as they are with someone who is suicidal, that is a good example of a time when 80% isn't good en…

> Most people who have worked with a professional therapist understand intuitively why the only helpful feedback from an LLM to someone who needs professional help is: get professional help. I'm not a therapist, but as I understand it most therapy isn't about suicide, and doesn't carry suicide risk. Most therapy is talking through problems, and helping the patient rewrite old memories and old beliefs using more helpf…

“If it is prompted well”

What the fuck does this even mean? How do you test or ensure it. Because based on the actual outcomes ChatGPT is 0-1 for preventing suicides (going as far as to outright encourage one).

Re: OpenAI says over a million people talk to ChatGPT about suicide weekly

#412

Post by mschuster91 »

basisword wrote: post #374
formerly_proven wrote: post #351

Earlier quoted context omitted.

> There are 24/7 suicide prevention hotlines at least in many countries in Europe as well as US states. My understanding is these will generally just send the cops after you if the operator concludes you are actually suicidal and not just looking for someone to talk to for free.

I mean that's clearly a good thing. If you are actually suicidal then you need someone to intervene. But there is a large gulf between depressed and suicidal and those phone lines can help without outside assistance in those cases.

> If you are actually suicidal then you need someone to intervene.

Yeah, trained medics, not "cops" that barely had a few weeks worth of training and only know how to operate guns.

Re: OpenAI says over a million people talk to ChatGPT about suicide weekly

#413

Post by sega_sai »

alex-moon wrote: post #317
Al-Khwarizmi wrote: post #236

Earlier quoted context omitted.

We would need the big picture, though... maybe it caused that death (which is awful) but it's also saving lives? If there are that many people confiding in it, I wouldn't be surprised if it actually prevents some suicides with encouraging comments, and that's not going to make the news. Before declaring that it shouldn't be near anyone with psychological issues, someone in the relevant field should study whether the…

I suspect you've never done therapy yourself. Most people who have worked with a professional therapist understand intuitively why the only helpful feedback from an LLM to someone who needs professional help is: get professional help. AIs are really good at doing something to about 80%. When the stakes are life or death, as they are with someone who is suicidal, that is a good example of a time when 80% isn't good en…

If I had to guess (I don't know) the absolute majority of people considering suicide never go to a therapist. Thus while I absolutely agree that therapist is better than AI, but the question is whether 95% of people not doing therapy + 5% people doing therapy is better or not than 50% not doing therapy, 45% using AI, 5% doing therapy. I don't know the answer to this question.

Re: OpenAI says over a million people talk to ChatGPT about suicide weekly

#414

Post by lazide »

mapt wrote: post #409
fragmede wrote: post #367

Earlier quoted context omitted.

Plastic bags shouldn't be allowed near suicidal teens. Scarves shouldn't be. Underwear is also a strangulation hazard for the truly desperate. Anything long sleeved even. Knives of any kind, including butter. Cars, obviously. Bleach is the least of your problems.

We have established that suicidal people should be held naked (or with an apron) in solitary isolation in a padded white room and saddled with medical bills larger than a four-year college tuition. That'll help'em. One problem with treatment modalities is that they ignore material conditions and treat everything as dysfunction. Lots of people are looking for a way out not because of some kind of physiological clinica…

The one dude that used the money to build a self-murder machine and then televised it would ruin it for everyone though. :s

The reality is most systems are designed to cover asses more than meet needs, because systems get abused a lot - by many different definitions, including being used as scapegoats by bad actors.

Re: OpenAI says over a million people talk to ChatGPT about suicide weekly

#415

Post by Dilettante_ »

ml-anon wrote: post #394
Dilettante_ wrote: post #359

Earlier quoted context omitted.

Can you give just a little bit more effort explaining why you say that?

You know, usually it’s positive claims which are supposed to be substantiated, such as the claim that “LLMs can be good at therapy”. Holy shit, this thread is insane.

I wasn't saying your position is wrong, just that it doesn't really make a good contribution to the discussion.

Re: OpenAI says over a million people talk to ChatGPT about suicide weekly

#416

Post by nelox »

spookie wrote: post #353
ethbr1 wrote: post #333

Earlier quoted context omitted.

You're holding up a perfect status quo that doesn't correspond to reality. Countries vary, but in the US and many places there's a shortage of quality therapists. Thus for many people the actual options are {no therapy} and {LLM therapy}. > This is why we have safety regulations, why we don't let people drive cars that haven't passed tests, build with materials that haven't passed tests, eat food that hasn't passed t…

Ok then, the LLMs must pass the same tests and be as regulated as therapists. After all, it should be easy peasy (:

What tests? The term “therapist” is not protected in most jurisdictions. No regulation required. Almost anyone can call themselves a therapist.

Re: OpenAI says over a million people talk to ChatGPT about suicide weekly

#417

Post by torben-friis »

danaris wrote: post #386
torben-friis wrote: post #368

Earlier quoted context omitted.

>If it cannot be used safely by a vulnerable class of people, and identifying that class of person sufficiently to block use by them Should we stop selling kitchen knives, packs of cards or beer as well? This is not a new problem in society. >and its primary purpose is simply to bring OpenAI more profit This is true for any product, unless you mean that it has no other purpose, which is trivially contradicted by the…

There's a qualitative difference between knives and publicly-available LLMs. Knives aren't out there actively telling you "use me, slit those wrists, then it'll all be over".

I don’t disagree that they are clearly unhealthy for people who aren’t mentally well, I just differ on where the role of limiting access lies.

I think it’s up to the legal tutor or medical professionals to check that, and providers should at most be asked to comply with state restrictions, the same way addicts can figure on a list to ban access to a casino.

The alternative places openAI and others in the role of surveilling the population and deciding what’s acceptable, which IMO has been the big fuckup of social media regulation.

I do think there is an argument for how LLMs expose interaction - the friendliness that mimics human interaction should be changed for something less parasocial-friendly. More interactive Wikipedia and less intimate relationship.

Then again, the human-like behavior reinforces the fact that it’s faulty knowledge, and speaking in an authoritative manner might be more harmful during regular use.

Re: OpenAI says over a million people talk to ChatGPT about suicide weekly

#419

Post by tsegratis »

tim333 wrote: post #403
alex-moon wrote: post #317

Earlier quoted context omitted.

I suspect you've never done therapy yourself. Most people who have worked with a professional therapist understand intuitively why the only helpful feedback from an LLM to someone who needs professional help is: get professional help. AIs are really good at doing something to about 80%. When the stakes are life or death, as they are with someone who is suicidal, that is a good example of a time when 80% isn't good en…

I tried therapy once and it was terrible. The ones I got were based on some not very scientific stuff like Freudian and mostly just sat there and didn't say anything. At least with an LLM type therapist you could AB test different ones to see what was effective. It would be quite easy to give an LLM instructions to discourage suicide and get them to look on the bright side. In fact I made a "GPT" "relationship therap…

the 'therapist effect' says that therapy quality is largely independent of training

some research on this: https://psycnet.apa.org/doiLanding?doi=10.1037%2Ftep0000402 https://pmc.ncbi.nlm.nih.gov/articles/PMC8174802/

CBT (cognitive behavioural training) has been shown to be effective independent of which therapist does it. if CBT has a downside it is that it's a bit boring, and probably not as effective as a good therapist

--

so personally i would say the advice of passing on people to therapists is largely unsupported: if you're that person's friend and you care about them; then be open, and show that care. that care can also mean taking them to a therapist, that is okay

Re: OpenAI says over a million people talk to ChatGPT about suicide weekly

#420

Post by jack_tripper »

kakacik wrote: post #378
jack_tripper wrote: post #354

Earlier quoted context omitted.

>Countries vary, but in the US and many places there's a shortage of quality therapists. Worse in my EU country. There's even a shortage of shitty therapists and doctors, let alone quality ones. It takes 6+ months to get an appointment for a 5 minute checkup at a poorly reviewed state funded therapist, while the good ones are either private or don't accept any new patients if they're on the public system. And ADHD di…

Not sure where you are based, but in general GPs shouldn't be doing psychological evaluation, period. I am in Europe, and this is the default. If you live in utter shithole (even if only health-care wise), move elsewhere if its important for you - it has never been easier, Europe is facing many issues and massive improvement of healthcare is not in the work pipeline, more like the opposite. You also don't expect butc…

>but in general GPs shouldn't be doing psychological evaluation, period. I am in Europe, and this is the default.

Where did I say GPs have to do that? In my example of my friend's being misdiagnosed by GPs, it was about another issue, not mental, but it has the same core problem of doctors misdiagnosing patients worse than a LLM bring into questions their competence or that of the health system in general if a LLM can do better than someone who spent 6+ years in med school and got a degree to be a licensed MD to treat people.

>You also don't expect butcher to fix your car, those are as close as above

You're making strawmen at this point. Such metaphors have no relevance to anything I said. Please review my comment through the lens of the clarifications I just made. Maybe the way I wrote it initially made it unclear.

>You get what you pay for at the end

The problem is the opposite, that you don't get what you pay for, if you're a higher than average earner. The more you work, the more taxes you pay, but get the same healthcare quality in return as unskilled laborer who is subsidized.

It's a bad reward structure to incentivize people to pay more of their taxes into the public system, compounded by the fact that government workers, civil servants, lawyers, architects, and other privileged employment classes of bureaucrats with strong unions, have their own separate heath insurance funds, that separate from the national public one that the unwashed masses working in the private sector have to use, so THEY do get what THEY pay for, but you don't.

So that's the problem with state run systems, just like you said about corruption, that giving the government unchecked power over large amounts of people's taxes, allow the government to manipulate the market and choosing winners and losers based on political favoritism and not on the fair free market of who pays the most into the system.

Maybe Switzerland managed to nail it with their individual private system, but I don't know enough to say for sure.

Post reply on HN