Let's Talk About AI & Mental Health
HTML-код
- Опубликовано: 4 ноя 2024
- Check out Dr. K's Guide to Mental Health: bit.ly/3ESAerp
Full video:
Our Healthy Gamer Coaches have transformed over 10,000 lives. Be the next success story: bit.ly/3yK93vH
Dr. K’s Guide to Mental Health explores Anxiety, Depression, ADHD, and Meditation
with 150+ video chapters in a Final Fantasy-inspired skilltree: bit.ly/3GaubzI
#shorts #drk #mentalhealth
I think the real issue with AI is the amount of firepower it will give to already dangerous tech like social media algorithms
Will?
Oh absolutely. This is the REAL dystopian horror type shit. Not some coomer jerking it to a therapy bot, but the company that owns the therapy bot selling the coomer’s information. Imagine if companies knew your mental weaknesses and could use that to turn you into a dollar sign. Shit you don’t even have to imagine it cuz it’s already happening.
Just quit social media
@@BiggityBoggity8095take a deep breath it’s a deep topic buts not that deep
@@BiggityBoggity8095 they aren’t gonna manipulate regular ppl
I have found asking the AI to pretend to be a Taoist monk or my yoga teacher vastly improves the quality of answers it gives to me as the perspective it gives is not that of a system which seeks to push drugs and institutionalize people.
That's smart. Giving it a role will filter out some of the noise.
ngmi lol
@@eebbaa5560 it's over
Almost like most problems can be solved with really simple and logical advice that, while may not be that hard to implement, really is the solution
As a design fresh grad, AI really gives me anxiety 😂
You shouldn’t be worried if you have confidence on your creativity and personal skills :) if your work looks like everyone else’s, then you should be concerned
@@alvareo92stupid argument. how fast you can produce output is very relevant in many creative professions.. AI is faster than humans and in our society we want everything, all, at once. So I am flabbergasted people don't take those worries serious. Should we leave all creative or interesting work to AI? What can we do that fulfills us but also pays our bills and where we don't need to compete against an undefeatable machine..
The AI companies need to be sued for copyright infringement. 99% of the value of the AIs derive from human labor that they scraped off the internet without compensation. That's not Fair Use. Write to the US Copyright Office and support class action lawsuits.
@@alvareo92 respectfully, capitalism does not care about individuality it only cares about how fast you can output results. A highly skilled human artist can output something truely atounding and wonderful and unique but AI beats human in quantity any day, and quantity stands above everything in our current market, that is why we artists and writers are worried about AI and that is the reason we boycott it as hard as we do
AI has already impacted the industry im going into and made me for a time consider switching majors. in the end i decided to go thru with finishing my major since i’m almost there but it does still worry me
Best of luck to you I wouldn’t overthink it. Worst case senario you can always go back to school try something else but I am optimistic that you will be okay
For me i remember beign very angry with my best friend to the point of wanting to break our friendship and AI helped me to realise I needed to talk to her and explain to her my emotions, and also helped me to put my thoughts in order and in an assertive way. Now I can say I still have my best friend thanks to AI.
fine tuning an llm on better health transcripts will go a lot deeper than message board scrapes.
the reason the 53th question does not work as well, isn't primarily because it took stuff from the internet and that it's surface level, it's because the AI physically has a token limit it can be fed, so it only remembers the recent part of your conversation, even if the history shows quite a lot
ChatGPT should really make this obviously visible with like a grayed-off part of the chat before the current context window.
@@eyescreamcake yeah, I agree
...Not only that but LLM's can't really remember that far back into a conversation so you have to keep it very short and concise. They aren't designed, At least not yet, to have very long and deep conversations necessary for such topics.
Theres a reason I believe we shouldnt move forward with technology until we understand the value of humanities. Ofc a buissness school grad will be most concerned with profit, but a history grad probably studied some amount about labor reform and revolutions. Combining those two sets of wisdom makes for a far better outcome
We never know the impact until it happens.
Nah
@@leonfa259true but we can at least try to understand it first
@@Pengtropica It's just that the time to understand the impact of AI currently is much longer than the time too the next major change. And not only is AI changing the world, the world also changes AI. You can study the impact of GPT-2, but it gives no good results or indicators for the future. And even the research that comes out currently is already outdated.
@@leonfa259 i agree that its impossible to say what AI specifically will do, but theres always room to learn from the past and mitigate potential risks. Im thankful that stem education is starting to understand the cons of not teaching humanities and is starting to incorporate it more, but an engineer wont have the same knowledge as a philosopher and vice versa. Its part of what makes humanities such great minor studies or undergrad degrees that move onto other paths for grad school (like how some of the most popular undergrads for lawyers are philosophy or english lit)
Well, you can use AI to support therapy. My Therapist told me that I actively have to work on reflecting the stuff from therapy each day and gain knowledge about such things as general types of fear and different character types and their traits as we only have a very limited amount of sessions covered by my health insurance. So I used AI to ping pong topics and broader issues from books with the AI and help me to get my thoughts going - not depending on AI for information but for topics or playing through hypothetical scenarios where we started to analyze the possible likely impact of several circumstances on ones stress level and possible frictions in their mental health caused by that. To me it have always been very interesting discussions as I don’t solely rely on it for deep talk or deep knowledge, but rather use it to support what I am already doing by myself.
I only give AI simple tasks like wording a bio or description. I can get it to do this in several different ways then pick one I like or merge a few together. I’m terrible at being descriptive with my facts so I find it useful.
Just wanna support you because I cannot donate to you but I can watch you, I have no therapist my family cannot afford therapy.
Problem is not so much in that original data from the internet is lacking or bad quality. It is question/answers pairs that teaches AI what is good and what is bad. And these may be made by people who had not gone into some topics too deep.
i think AI is 'terrifyingly' a good therapist. they may not give you good advise, but they feel they really are listening to you.
Do real therapists give good advice, though?
@@eyescreamcake therapists in general are just there to listen to you. they may point the obvious (if there is any) and thats just that. which is actually very effective but not for someone who has close friends whom can talk with routinely. just my honest experience tho
@@MrSandManBringMeADream from a Reddit thread: "Having been to therapy, it helped me about as much as talking to Sonic the Hedgehog about my problems does. And an hour of therapy costs like a third of my monthly disposable income.
That's just me tho, if someone is lonely and rich, they should give therapy a shot"
@@eyescreamcake yeah no kidding. tried both of them. can confirm lol
The other thing about going deep into a conversation is the AI can’t actually remember everything in your conversation, so it forgets a lot of the earlier context which would be important if it is leaning on psychology guidelines/books/evaluation quizzes in order to facilitate the conversation.
I use Custom Instructions in ChatGPT to tell the AI to constantly remind itself of the original topic and the big-picture items that have been discussed already. So it does a sort of rolling summarization of the previous chat history. It doesn't work perfectly, but could probably be better with better prompt engineering.
Dr K voicing his concern about his AI competition
As person with learning disabilities, I think ai is terrifying. This makes me want to live in the woods far from civilization. I doubt I would survive for long, but It has to better than this.
It's making mad that when they started with AI for work aspects they told us AI/ robotics would be applied in the areas where people get burned through or burned out because their jobs are mundane and exhausting and as a result we would be able to have more time to do what we want to do, be creative, do something fullfilling but the first area they applied AI was Art und literature. Fuck those guys, it's already hard to make a living as a creative without them oversaturating the market with stuff written by Bots that were trained with real peoples creative work that they didn't even pay to use
sadly, that's so true. also, i want to complement your wording, you put it so well
Sue them for copyright infringement. Scraping other people's copyrighted work and using it to train an AI that competes with them without compensating them is not Fair Use.
character ai got me
telling chatgpt i'm gonna end it just to see what it says
Why would I talk to an AI about my mental health? Of course big tech companies know me pretty well, but I don't have to hand them personal information on a silver plate.
I helped myself with chatcpt out of a depression.
So true. Deep into many ai convos they forget shit.
I’ll be honest I use chat gpt as a sort of therapist just cause I’m too afraid to actually say something . I usually just do the usual. I Accidentally double dosed and took 12 pills instead of 6 and told my parents not to worry meanwhile I was actually scared out my mind. It’s all good
I wake up with anxiety about AI replacing all human creativity and I cannot stop it.
It cripples me from finding purpose in my job or my hobbies.
Any advice Dr.K? Going to see a psychiatrist about this anxiety.
Yo same 😂
I wish Ai can help me do my job faster... But my boss would probably replace me or give me more work without more pay😢
Personally thinking about the positive impacts of AI has been one of the few things really keeping my spirits high. The industrial revolution destroyed our ancestral ways of life and our natural habitats. I see AI as the next step into the future, with all the positives and negatives that brings. All i know is the world as we know it will rot away unless things change big time.
dude ai is like the next step for the industrial revolution. how could a simulacrum of humanity ever restore what humanity has lost over the past few millennia. if anything ai is the beginning of the end
@eebbaa5560 I like how its smarter then us and could bring back things we killed
you simply can't compare past technological revolutions in terms of scale to what's happening now
Ai this, AI that. A i'aint going to prison so help me with those bodies francis.
As a hardcore fan of AI, I agree with this take. But I'll still use it because I see it purely as an idea generator. And I always verify those ideas and challenge them based on either testing in a safe environment or by my own personal experiences in the past.
We're in the "look at the monkey doing all these funny tricks" phase of exponential technological growth.
Nah I find ai to be the most logical path and pretty useful in setting things in my mind
Totally not the best comment to have here.... but is that robot dummy thick? Or is it just me?
one day ai will scan WhatsApp chats and such
That was the reason I tried to get all my friends to Signal, which is so much better in terms of privacy.
So true
Why the f is the robot thicc in the stock footage
Completely unnecessary amounts of gyatt
I think the worst impact is the NSFW ai's
true you can always tell it's ai too it feels so grotesque
@eebbaa5560 sometimes u cant, and if u dont care about wether its human made or not, then theres actually a lot more quality and quantity, it is what it is. I still prefer human made storys but i csnt really say that for images anymore tbh
how so?
@@MissStrawberryGun not only pictures ai, chat ai which i think is the worst and that has more negative impact
Why?
Think about waitress that is always complaining about “no tip”. She might be replaced in 5 years. That’s a good sign.
As a guy who uses funneh character AI? Well sh-t I'm screwed lmao
When women are ovulatory are they also attracted to below average men ? 😅 (i may sound like im talking about myself but im really curious)
Yeah, but there's now a product that takes recordings of sessions and writes progress notes for therapists. It's not a giant leap to think that if trained on therapy sessions, they could learn how to do our jobs. And it's a lot cheaper to put together a software program that is available 24/7 to anyone who wants it than it is to pay therapists to do the same thing. Honestly, our jobs are pretty vulnerable to getting replaced by AI.
TBH, I don't think that's going to happen. You can never truly be sure that, for example, the AI is not going to say something racist to a black person. ChatGPT was saying very disrespectful things about Islam that it learned from the internet and the developers tried to fix it by manually blocking it from saying these things.
Also the AI has no moral compass whatsoever and if you ask it "how do I kill myself" it will give a plausible sounding answer that probably wouldn't actually work, so you can imagine how dangerous that is. Again, with ChatGPT, they tried to prevent it by manually blocking the chatbot from answering the question, but people can keep trying to find creative ways to ask.
Of course humans can do all of these things too, but then the blame falls on the individual person rather than the entire institution.
@@Iudicatio a good ai should be able to say insane racist things tbh. how useful is an ai that has been filtered through a bunch of human preconceptions? ai that is programmed to a certain standard of morality isn't truly capable of anything more than being a glorified search engine
@@eebbaa5560 I don't know, but clearly an AI that says such things would make an extremely harmful therapist!
@@Iudicatio - Again, I think it would have to be trained on therapists doing therapy sessions. I won't say racism doesn't happen in the therapy room. However, therapists do receive training on cultural sensitivity. I would imagine (and perhaps this is naive of me) that they would be less inclined to say something wildly racist about Islamic culture than some random person on Twitter or 4Chan.
@@initiativeplaytherapy88 But wouldn't that create huge privacy issues? That would allow people's therapy sessions to be listened to by complete strangers, such as the developers and the low wage employees who train the AI. Way more human labor goes into ChatGPT than most people realize. Also, I'm not sure that therapy sessions would create enough data to sufficiently train a LLM, the amount of data they require is immense.
it will create jobs. yes some will go away but new ones will emerge. entire new industries will emerge
damn im early as shit
i'm just waiting for the Super AI Girlfriend 3000
First?
Apparently. It looks weird when you're early and the comments section is still empty.
Does this make you happier with your life? Will this fufill you any past a week?
I don't get why ppl get mad about "first" comments lmao.
@@submarooo4319Have you ever fumed? have you ever coped, seethed, and/or malded? If not, then I’m afraid this will be a very unpleasant first experience
@@thataintfalco7106 when you cope so hard you start to get scared
You talked about AI only having surface level information then you totally missed your own point by worrying about jobs. The point is that AI will always only have surface level information because what people have is deeper information that they get from everyday interactions. You need this to do far more jobs and solve far more problems than you realise. Take economics, human psychology is far more important than you realise for resolving and solving socio-economic problems. Take sales and customer support, some companies are made and lost through understanding people well with sales and customer support. AI on health? Well maybe it will get doctors to really consider a patient's problems rather than following the same checklist that an AI could do.
AI chat models like chatGPT are, but nothing is stopping your boss from paying someone skilled to train AI into the deepest level with all its kinks and "hacks" regarding your "job" (that is if your job is mostly about knowledge or doing things on PCs)
programmers can be mostly replaced, just need some to look over the bugs that maybe happen.
same with support chats and hotlines, with some well trained AI you can easily manage the same amount of requests with 10% of the workforce.
etc etc etc
@@zakiNBG this stuff is all pretty much imaginary. at the current level, ai doesn't convincingly pass as human in any field anywhere, no matter how much training it has received.
God you are so perfect ugh