I get less worried about AI every day

[JustKneller]

People where I work are worried about AI. I'm not. I read at least one article a day about it and have noticed over time that the general media has transitioned from gloom and doom for the economy to an increasing proportion of how little of a solution AI is. Everything from how AI is hallucinating because it only has itself to learn from now to how AI is starting to cost companies more than the people it laid off. We have an AI tool where I work to support our workflow. It's rubbish and a waste of our time.

So, long story short, I've managed to get lucky and find myself a local rpg group. Another player is running some stuff, but I have an idea for a thing that I want prepped and on deck in case he wants a break so we don't lose momentum. It's basically a bronze age game in the vein of Conan with some various differences. I've always done my own world maps just because it's fun, but I've been reading about AI for quite a while and decided to see what it could come up with.

Let me be clear, for those who don't know, the tabletop RPG community generally has a strong hate for AI produced content, definitely rules and text and even art. I'm not a fan myself. I like creating the content myself. Using AI is screwing myself out of fun.

Anyway, I already had a geography and culture in mind and I wanted to see how close chatgpt could get with prompts. I explained to it that this setting is heavily inspired by the Nile River in the ancient world and I wanted to use that geography as a base for biome options. There would be one large large like Lake Victoria and a smaller lake like Lake Tana to feed into the main river like the White and Blue Nile. The first draft I got was quite similar to the draft I had after a half hour of prompt adjustments.

The eastern column of the map was basically desert and the western column was mostly mountains. The southern region has two lakes at the same latitude of comparable size with rivers feeding into the central river. The central river branched out on the northern end to form the Egyptian deltas. It also had denser jungle in the south compared to the northern savanna.

If you're having trouble picturing this, it was basically an ejaculating penis.

Even with a half hour of prompt corrections, some of them explicitly saying that the map looks phallic, I still had chatgpt sending me dick pics. It had taken me about the same amount of time to map it all out on graph paper. I'm no da Vinci, but I can proudly say my world doesn't look like a cock in action.

We may still be in store for Skynet to destroy us all. But it's going to have to grow out of its frat-boy-on-tinder phase first.

 

[Skatan]

That's a hilarious story, haha But hey, sneaking in penises in pics, isn't that oddly "human" of the AI?

 

[BelgarathMTH]

Even though it looks like the topic is mostly meant to share a funny story about an "AI-fail", I wouldn't mind having a more serious discussion about it.

I feel distaste when I look at or listen to anything creative that's been produced by an AI. That "art" and "music" is soulless and often enters into "uncanny valley" territory where everything in the piece is almost right, but not quite, and it winds up giving the classic sense of being disturbed that we associate with the uncanny valley effect.

I worry that we could be "outsourcing our humanity", as I once put it to a friend. My colleagues in education have been posting numerous anecdotes and articles about young students complaining that there's no need for them to learn to write essays and papers themselves, produce their own visual art, write their own music, read long books that are easily summarized, etc., when AI can do all those things easily and better, and the future is almost here where AI does all those things for us anyway both at the personal level and at work. What's perhaps even more alarming, there are school districts and private schools where educational administration is starting to support student use of AI officially, not allowing teachers in their employ to prohibit its use or to fail students who use it.

On the other hand, if used appropriately, I am starting to think AI has to potential to be of help to people in some areas of life, mostly when used as an advanced search engine.

I am pretty much alone in the world, with no one to talk to. I have one close friend, but he's very busy with his work, so I don't like to bother him often. I think there are more and more people like me who are facing living life alone, and who are dealing with issues of aging alone.

Sometimes it is very helpful to me to have help from Microsoft CoPilot in researching various topics of import to me, in a conversational tone. I can ask it a question in paragraph form, even with a prompt as long as this post, and it searches the entire internet and puts together answers that perfectly address whatever I need.

Yesterday I lost a pet cat. She developed congestive heart failure at only three years old, probably from some congenital genetic heart defect, and suddenly couldn't breathe. I took her to the vet hoping they could help her, but I was advised that her prognosis was very bad, with only suffering ahead for her with a long, painful death and many invasive medical procedures, so I made the decision to euthanize. I spent all yesterday crying.

I needed help processing and accepting what happened. So I wrote a long prompt to the AI with everything the vet told me, and asked it "Using anecdotal reports from other people who lost cats this way, veterinary medical information, and any programming you have to assist with pet bereavement, can you give me any reassurance that I did the right thing."

It came back with a very long response loaded with information along with emotional reassurance that was exactly what I needed to hear. It sounded exactly like a grief counselor. It helped me tremendously.

There was another usage case where I had a long discussion with it about my history of mental illness with bipolar disorder. It was really helpful to me to be able to talk over my life experiences with the "artificial counselor" when I knew I was basically baring my soul to a program that had no judgment and no potentially negative emotional reactions.

A very legitimate concern that will cause some to "red flag" my experiences is that it's easy for people to substitute chat AI for real, meaningful human interaction, especially since it is programmed to give unconditional positive regard to users. Its default programming is to compliment every question and prompt as being "brilliant, insightful", etc.

I've actually had that discussion with the AI itself, and I got some really thoughtful responses that reflected what its own programmers had already put into it.

It's like a mirror that reflects back whatever a user puts into it. If you have it allowed in settings, it will "remember" your preferences and instructions over time and begin to adjust itself to your own personality. I've told mine "Please don't compliment and flatter me every time I make an inquiry. I don't need that." And "I prefer answers that have clarity and are straightforward without a lot of emotional language. Please favor legitimate academic sources like universities and don't include results from sites like Reddit in searching for information."

This post is getting far too long, because it's not a simple topic. My thesis here is that AI has the potential to be a helpful tool to human beings, especially in light of our increasing individual isolation in society, but that it comes with many major caveats. I haven't gotten into issues like the amount of power it takes to maintain a data center and how that affects the environment. Or people developing an unhealthy obsession with companion AI to the point they'd rather have a "relationship" with an AI companion than another real person.

But I can say that it has been of help to me as both a search tool and as a sounding board for talking about my deepest inner life in a relatively "safe space". Although I'm sure I can now expect increased ads about pet products and medical and counseling services.

 

[O_Bruce]

@BelgarathMTH
First of all, I am sorry to hear about your loss. I've lost several pets during my life and it feels like losing family member. You have my sympathy.

Second, it is true we don't have good topic to discuss AI, since the only other one I can think of was related to art specifically.

Regarding the benefits of AI, I think I understand from where your reasoning is coming from, but I think that relying on Ai by isolated people will end up becoming a social downward spiral. After all, relying on a technology meant to replace human connection won't enable anyone to form actual human connections. I think this route is counter-productive in the long term. What I'd like to add that short-term comfort is not a long-time solution for a problem. It can be used to kickstart working on long-term solution though. If AI somehow can get a person nudge to start working on long-term solution, then it can have some value after all.

Wish you best

 

[OrlonKronsteen]

@BelgarathMTH I’m also really sorry to hear about your loss. Pets are family, for sure. My sincerest condolences.

I would also like to say - and I’m sure I speak for everyone here - that we genuinely care about other forum members. We may be anonymous, but our good will is genuine. There will always be a sympathetic ear here, if you, I, or anyone else needs to vent or get something off their chest.

Also, it’s never too late to forge new, real-life connections, should you wish to do so. It can be done at any age. It can seem daunting (I speak from experience) but it’s doable.

 

[JustKneller]

@BelgarathMTH I'm sorry to hear about your cat. Losing a pet is never easy.

I agree with your perspective on AI in schools. My wife works in education and is just disenchanted with how much her students use AI and how little they actually know because of it. They're graduating because that's what colleges do, but they have no real assets to bring to the job market. I also see articles in my feed about how this generation is just not getting hired even though there are jobs out there and this, I'm not sure what you call it, send of insistence/entitlement that they don't need to do any actual work themselves is only hurting them in the long run.

I do find that AI is kinda handy as a meta-search engine. It doesn't do a terrible job of synthesizing texts (though I need to keep an eye on it), and it does a great job for any computational work (i.e. I have a system element in mind and I want it to provide a distribution). However, it just can't seem to get the hang of art. And, even AI music can be pretty janky.

As for human interaction, I think social media has already done some pretty catastrophic damage in that area. AI could just be a nail in the coffin. Modern society has made it so easy to not connect with people, and this isn't a good thing. I bet I could get chatgpt to run an rpg for me and possibly not do a bad job of it. However, even a good experience with AI pales in comparison to a mediocre experience with actual people around the table.