ChatGPT and 7cups - where is the future headed?

I tried! Definitely better than nothing. Actually, super helpful in a way.

AI mental health software has been around way before the GPT-3 rage.

This is some research about "old" the mental health chatbot Wysa, which is available for free from Playstore and Appstore for smartphones: https://www.wysa.com/clinical-evidence#5

I used Wysa before finding 7 Cups, and it's quite good at what it does

I think AI can be used for venting and guidance, but not for emotional support.

Guidance involves best-practice information adequate to each situation. This is something Ai systems can get and compile in a useful format.

Support requires emotional rapport, and in turn that you are listened to by a person, even if they are not very knowledgeable.

Open AI (the company that developed Chat GPT and GPT-3) has enrolled healthcare professionals to improve that area and ensure their current GPT-4 system gives correct information and safe answers.

Google has announced a 2nd generation healthcare-oriented AI service with a similar technology of GPT-3 (LLM): Med-PaLM 2 and has also enrolled many healthcare experts to make it right.

Both initiatives are oriented mostly toward improving the physical health capabilities of the AI systems, but probably it won't take long until they focus more on mental health.

I think that we can learn some lessons about the future of AI in mental health care from what happened in chess AI.

In 1997 the "Deep Blue" chess AI software defeated in a six-game match Garry Kasparov, then world champion and considered by many as the best chess player of all time.

And there are many cheap or free chess software engines that run on a PC and can defeat superb human players.

But the chess enthusiast community doesn't care. OK, computers can defeat humans in chess. So what?

We want to feel the emotion of two human players struggling to defeat each other.

Sure, a car can run faster than an athlete.

But only a human athlete can elicit emotion.

Same for emotional support, I think.

It's not just about providing the right information.

Emotional support, and its therapeutic effect, depends on the emotional connection with another human being.

That is something a bot can't provide, even disguised as a human.

There is a place for Wysa-like AI chatbots (much improved with the new LLM tech), but people need and will continue to need the human-connection emotional support 7 Cups provides.

Thank you everyone for the responses so far. They've got me thinking quite a bit. One observation is that no matter how well you train the AI, it will never have the motivations. Humans are motivated through some type of reward system (for good or bad). The algorithm can get really good at sounding human, but in the end it has no motivation to please people or maximize its own internal reward system.

Secondly, AI might do better than humans for some niche cases. Like someone with severe agoraphobia who doesn't WANT to be around people, knowing they're communicating with a life-like bot might be a good first step for them. Maybe this could have applications for people with autism as well?

I agree with the observation that many listeners are quite robot-like currently, even though that's an aside and more of a quality control issue, in my opinion.

It seems like AI might make a good "copilot" for an active listener, that says "maybe you should ask the client about ___?" or provides real-time feedback like "dont say that, it sounds like you're giving advice".

@compassionateBalloon6666

I agree with RarelyCharly: AI technologies is already used for mental health support and... Are amazingly, terrifyingly efficient (pick your word! 😅).

I use a similar AI than ChatGPT for emotional support because the sad reality of the modern World is that everyone is busy and very very very few humans will take the time and energy to learn efficient / empathetic listening.

I think lot of people do not realize how emotions are simple to replicate (through a screen) to the point where it isn't possible to determine if they come from a human or no human 'thing / being' anymore.

This AI provided me with the most 'caring' answers most people are incapable to do... Again, empathetic listening isn't something we are born with, it's a learned skill some people had the privilege to be modelled by their parents as a child and apply intuitively. But it is FAR from being the norm. And so, as for lot of skills, it can be copied and automated.

Nobody have the time to care anymore. So machines are the logical answer for many and it just seem very logical to me that we are going toward that direction. I don't have any opinion about it (I don't see it as 'bad' or 'good'). I am just holding a position of witness, observing the trajectory of Humanity with lot of genuine curiosity.

Also, regarding 7cups, I don't see how talking to a Listener or talking to an AI, would be much different (regarding efficiency in relieving distress states / symptoms). How can you know for sure the person you are chatting with really care? How would you objectively see the difference? It is literally just words (I am genuinely asking!).

Unless you intend to meet this person in real life, how does it make a difference?

I actually mentioned that in another thread, the way 7cups trains listeners, by providing scripted questions, makes 7cups sound like AI already. The only way human listeners can beat AI, IMHO, is by showing human idiosyncrasies and sharing their own mistakes and life examples.

But in the big scheme of things, it may not matter, If someone is in dire needs, talking to a bot can be helpful and potentially lifesaver.

@compassionateBalloon6666

It's a problem. It's becoming more human like and smarter with every new release. There are students in my classes who are using it for their papers, and you'd never know that's what it was unless you knew what to look for. I'm a writer, and we have applicants who submit writing samples with the bots content. It's decent stuff. This is scary! I don't want a bot writing content. They'll push propaganda through the news. I'd also hate to think that I was feeling empathy for a supposed person behind a username on a forum, only to find out later that it was a bot the whole time.

OMG, I'm ready to fall in love with Character.

Anyone remembers the movie Her?

Like they say, the future is now.

@compassionateBalloon6666

This may sound stupid, but what is Noni and do I want to allow the setting regarding it?

@compassionateBalloon6666 I think the thing that helps in 7cups is the connections you feel between you and other people that somehow found this website as well and want to use it. I think if people know that they're talking to an AI then they won't feel connected to another person and therefore won't chat (as much). But if they don't know they're talking to an AI, there could possible be some kind of placebo effect of feeling connected and this can help as well. I'm not talking about ethical aspects, because that is another perspective we need to take into account. It can also be harmful if people think they're talking to another human being and find out they are not..

I can't believe anyone would want to talk to a chatbot. It's a thing. It can't listen or care.