Artificial Intelligence

[Psy-Fi]

Put a bunch of AI-driven robots in football (American or European) or the Olympics, and how fun and enjoyable would that be to watch?

Probably wouldn't hold my interest beyond a single game but I gotta admit that's something I'd like to see.

[SGR]

The whole point of art is that it's not like a sport at all though. It's not about that kind of achievement, it's about channeling some of your inner personal world into an art form so other people can try to access it. And about making beautiful things too.

So that means that I think if an AI could do that and we disregard environmental qualms, then there would still be the obstacle of not connecting with another human being; otherwise it would be fine. AI is just not nearly good enough to do that, and because a particular kind of originality would be required, I think it's pretty far away. It can just make things that look or sound good, that's nothing special, it's not art

It's definitely an interesting discussion - I might argue that though there are very obvious differences (one of the most important distinctions in my eyes is that art broadly does not have rules, while sports broadly do have rules), art is in some ways like a sport in the sense that the value we place on it, at least up til the present day, is very much tied to the human element. Humans are the throughline, and in both endeavors, mind and body together first must conceptualize (for art: creative idea/form of expression/medium, for sports: deciding to practice, planning a routine, etc), and then actualize (for art: the physical act of creation, for sports: training, competing, taking supplements, etc).

To your points about where we're at with AI 'art' now, I don't know that any kind of originality is required for something to be qualified as 'art'. But, Oxford defines 'art' as the following:

the use of the imagination to express ideas or feelings, particularly in painting, drawing or sculpture

Where we are today, I don't think what AI is doing could qualify as 'imagination' by any definition of the word - so that's what would disqualify its creations as art in any strict/technical sense. If we're talking layman's terms though, what it creates is art in certain a way, specifically if a human has used the AI creations to buttress their own art - in other words, in isolation or in a vacuum, what AI creates is not art, but once a human uses the creation for their own creative endeavors, it is art (holistically or as a piece of the whole: as the human imagination was involved in using the tool, similar to how a painter might use a specific paintbrush, or a specific paint).

This whole distinction could be turned on its head if we ever get to the point of AGI (Artificial General Intelligence), where the machines can learn on their own and replicate the abilities of human intelligence - if we ever are to get there, the distinctions between a human imagination and an AI "imagination" might be so blurry as to seem non-existent.

[SGR]

Probably wouldn't hold my interest beyond a single game but I gotta admit that's something I'd like to see.

It would be novel, I can't argue that. 

My guess though is, like you said, it wouldn't retain interest like professional (human) sports leagues do today.

[Psy-Fi]

It would be novel, I can't argue that. 

My guess though is, like you said, it wouldn't retain interest like professional (human) sports leagues do today.

The Jetsons Football 🏈

[Trollheart]

I think the real litmus test would be to say to an AI program "Create a work of art but do not refer to or use any other artist's work, don't base it on anything already created; make something entirely original that only exists in your own imagination." I doubt it could do it, because as you say, SGR, AI really at the moment can only replicate. You can ask a Chatbot to write a story in the style of Dickens or Tolstoy or anyone else, and it can. But ask it to create an entirely original story without referencing any other works, I don't see how it could.

[SGR]

I think the real litmus test would be to say to an AI program "Create a work of art but do not refer to or use any other artist's work, don't base it on anything already created; make something entirely original that only exists in your own imagination." I doubt it could do it, because as you say, SGR, AI really at the moment can only replicate. You can ask a Chatbot to write a story in the style of Dickens or Tolstoy or anyone else, and it can. But ask it to create an entirely original story without referencing any other works, I don't see how it could.

You are quite right that an AI would be unable to truly accomplish that task as things stand today. And though right now I agree that the concept of 'imagination' and its uniquely human employment to create expressions is what seems to technically distinguish human-created 'art' from AI-created 'slop', even we humans with our seemingly unique 'imaginations' can hardly create something original wholecloth - we too in our brains and imaginations are certainly influenced by many different inputs in creating our outputs: the painter perhaps by Da Vinci, Picasso, and Van Gogh - the author perhaps by Hemingway, Tolstoy, and Dickens. If one were a certified expert in either field, my presumption would be that one could look at or read a painting or a novel and make a pretty good guess as to the influences of the creator. But our human outputs are not created in a very strict, cold, machine-like way based on the input - we are capable of using those influences to create something that on the surface appears original and perhaps unique, even if the primary influences could be identified.

That being said, if we are to reach the point of AGI, it may be able take influence from hundreds of thousands of different inputs, varied, disparate, and perhaps even contradictory all at once, and use them to create something that appears to our human brains as something so complicated and convoluted in its possible influences, we can't even readily identify what the influences might be. As a result, its work may appear so original and impressive, we don't even think a human being would be capable of creating it. It might be beyond human appraisal.

It seems to be the human condition throughout our existence, as the social creatures we are, to build connections to each other - both physical and non-physical. In a way, it's almost like we're all individually just one cell in a larger more complex multi-cellular organism that we call humankind - sometimes working with each other, sometimes working against each other, but always connecting. Our efforts to build AGI may be our attempt to leverage all the knowledge we've gained through our connections over hundreds of thousands of years to rebuild God.

[Psy-Fi]

China races robots in Beijing half marathon

Doesn't look like they're quite ready to take on the humans. One of them looked like it was having some type of "health event."

[Buck_Mulligan]

@SGR & @Trollheart above, I think the distinction you're discussing is between Generative AI, which we have now, and AGI, which is definitely coming. Best I read is that they are not all that far away from AGI now.

[SGR]

@SGR & @Trollheart above, I think the distinction you're discussing is between Generative AI, which we have now, and AGI, which is definitely coming. Best I read is that they are not all that far away from AGI now.

They've been saying that for the better part of a decade now. We'll see. 

[Buck_Mulligan]

They've been saying that for the better part of a decade now. We'll see. 

Open AI's 03 released 04/16, although not exactly AGI, is apparently a big step forward...

Wiki:

o3 demonstrates significantly better performance than o1 on complex tasks, including coding, mathematics, and science.[1] OpenAI reported that o3 achieved a score of 87.7% on the GPQA Diamond benchmark, which contains expert-level science questions not publicly available online.[14]

On SWE-bench Verified, a software engineering benchmark assessing the ability to solve real GitHub issues, o3 scored 71.7%, compared to 48.9% for o1. On Codeforces, o3 reached an Elo score of 2727, whereas o1 scored 1891.[14]

On the Abstraction and Reasoning Corpus for Artificial General Intelligence (ARC-AGI) benchmark, which evaluates an AI's ability to handle new logical and skill acquisition problems, o3 attained three times the accuracy of o1.[1][15]

https://en.wikipedia.org/wiki/OpenAI_o3#:~:text=It%20is%20designed%20to%20devote,by%20o3%20and%20o4%2Dmini.

[DeepMinder]