Thank you for this fascinating and daunting post. I don’t think we humans have figured out what creativity is and where it comes from. We think we know it when we see/hear/experience creativity, but isn’t our individual creativity some form of patterning from our experiences and the data in our memory? Our originality comes from images, facts and chemical reactions. That might be all something AI can simulate. If so, AI will find more and more complicated ways to pattern its data and it will learn what we like because we will reward it. How many of us really know the difference between darn good and genius?
Great comment! Agreed, we don't know exactly what creativity is and where it comes from — and then there's the great mystery of why some people are so much more creative than others. What gives us a Mozart, a Picasso, or a Shakespeare? I do think we rely tremendously on the "I'll know it when I see it" test ... but then again, some people were only belatedly and posthumously recognized for their genuis. Much to ponder!
I think there are layers or levels to creativity. Or perhaps "expertise" is a better word.
As a personal example, I can't carry a tune in a bucket and I don't hear what musicians hear in music (which is why I don't like most jazz). An AI would help me, today, if I wanted to compose and produce a love song. Compared to a love song composed by Beethoven, it would be musical garbage. Compared to my own efforts, it would be amazing. For many cases, mediocre "creativity" is sufficient because in half of cases an individual human is below average.
How good is your radiologist?
These discussions always flounder on definitions. Is it "interesting" that a machine can beat the best human chess player? Maybe, but not very. The best human chess players are extreme outliers of any definition of "human". Anyone can learn the rules. Not everyone can "play" chess, let alone play well. Beethoven (and his few peers) are equally outliers. Attributing the creativity and genius of those few to humanity in general is a stretch, probably an unwarranted one.
For the near term, I have no doubt we'll get AIs that can trounce the average human at any particular endeavor. Some of these exist, now. Will we ever have a single AI that can trounce the best of the best humans as a group in all endeavors? I doubt it - but it hardly matters because the best of the best humans never act as a group.
See 'Virtuoso' in the story collection that I reviewed here:
https://chicagoboyz.net/archives/68808.html
Thank you for this fascinating and daunting post. I don’t think we humans have figured out what creativity is and where it comes from. We think we know it when we see/hear/experience creativity, but isn’t our individual creativity some form of patterning from our experiences and the data in our memory? Our originality comes from images, facts and chemical reactions. That might be all something AI can simulate. If so, AI will find more and more complicated ways to pattern its data and it will learn what we like because we will reward it. How many of us really know the difference between darn good and genius?
Is HAL inevitable?
Great comment! Agreed, we don't know exactly what creativity is and where it comes from — and then there's the great mystery of why some people are so much more creative than others. What gives us a Mozart, a Picasso, or a Shakespeare? I do think we rely tremendously on the "I'll know it when I see it" test ... but then again, some people were only belatedly and posthumously recognized for their genuis. Much to ponder!
I think there are layers or levels to creativity. Or perhaps "expertise" is a better word.
As a personal example, I can't carry a tune in a bucket and I don't hear what musicians hear in music (which is why I don't like most jazz). An AI would help me, today, if I wanted to compose and produce a love song. Compared to a love song composed by Beethoven, it would be musical garbage. Compared to my own efforts, it would be amazing. For many cases, mediocre "creativity" is sufficient because in half of cases an individual human is below average.
How good is your radiologist?
These discussions always flounder on definitions. Is it "interesting" that a machine can beat the best human chess player? Maybe, but not very. The best human chess players are extreme outliers of any definition of "human". Anyone can learn the rules. Not everyone can "play" chess, let alone play well. Beethoven (and his few peers) are equally outliers. Attributing the creativity and genius of those few to humanity in general is a stretch, probably an unwarranted one.
For the near term, I have no doubt we'll get AIs that can trounce the average human at any particular endeavor. Some of these exist, now. Will we ever have a single AI that can trounce the best of the best humans as a group in all endeavors? I doubt it - but it hardly matters because the best of the best humans never act as a group.