[Long comment, feel no obligation to read it, though at least I can assure you it wasn't AI generated!]
It's incredibly easy to take a photograph these days. At least one that only a small handful of people would care about. It's much harder to take one that is in some sense attractive to a larger number of people. And it's incredibly difficult to take one that is beautiful, that takes your breath away, that causes you to study it for minutes or longer.
I can't think of anything AI could do to diminish the first of these types of photos. People taking selfies, or pictures of their families and friends, or even of their food - an AI image won't diminish these in any way, and would likely be frowned upon if it tried. That's not *me* standing in front of this particular structure or landscape, those aren't *my* friends, that's not the meal *I* ate. We might have AI generate them anyway, just for the novelty of making me and my friends look like superheroes, or making it look like me eating an 18" tall hamburger, but we'll always know they're not real.
I also don't think the last category, Ansel Adams photographs, for example, also won't be seriously diminished by AI generated images for the same essential reason: they might look pretty, be perfectly composed, etc., but they're not "real" in the sense that they don't reflect something someone saw in the real world. Here it's possible that AI might financially impact the livelihood of a photographer in that you could tell it to generate an image that resembles something like theirs, but I feel that the impact would be moderate at best. People will either know it's not real, not really taken by their favorite super-star photographer, or if it has been claimed that it was, it will ruin the reputation of the publisher when it inevitably is pointed out that it's an AI fake.
So that leaves the middle group. Let's call them professional photographers. Many of them probably take photos for a living. Photos that fill our books, magazines, catalogs, games, web pages, and backgrounds, for example. It was these various industries that created these jobs in the first place. These are the people that will be most impacted by AI generated images, and it's simply an economic thing - the people paying them would likely take *any* cost savings they can get, not simply AI that's now able to produce images that are similar enough that no one really cares. This has been the case since the industrial revolution started. I don't think there's anything that can be done for it because if one company/publisher for some reason doesn't join the AI bandwagon, while most everyone else does, eventually they will be left behind. Maybe they'll have some niche cache amongst a small group that insist their buggy whips be hand-made, but sales will be minimal, because in general, no one likes to pay more for something when they can't distinguish a difference in quality just because it saved some unidentified person's job. They have other expenses and possibly a diminishing income as well.
Other art will likely follow the same pattern. I'm not going to stop doodling on my coffee cups just because an AI can do it better. And super-star artists will continue to work in whatever medium they are known for, and there will be little impact in the price they can get for their painting/sculpture/movie, or whatever. The big problem here is that much of the middle-group's art is digital. (That portion that isn't, probably is safe for a while yet. At least until robots get a lot better.) Digital copies don't cost anything, whether they are literal bit-wise copies, or knock-off imitations in the style of some artist created by an AI. And again, it is the publishers that created the market for these sorts of art in the first place that will eventually be forced by economics to reduce what they pay for it. Except for niche cases, if they don't, they'll simply be priced out of business, and the artists will be out of work anyway. I feel sorry that these artists will have to find other ways to make a living. They work in a field where their product has been commoditized and better ways of producing it are now coming available. I similarly felt sorry for all the machinists that lost their jobs because industrial robots now do it better, or because their jobs were shifted to a cheaper country where, if not yet, soon robots will do it better. Soon I suspect I will feel sorry for all the doctors that will lose their jobs as AI starts doing a better job than they do. (I won't feel very sorry for the lawyers who lose their jobs, though =) And I feel sorry for all the people that work in the programming field because they were told it was a lucrative career. I'll be looking forward to hearing your take on this, and how it compares to artists' losses.
But all our feeling sorry for these people never brought any of their jobs back. We could try throwing our metaphorical wooden sabots into the metaphorical gears of AI data centers, but it's not going to make a difference long term. Personally, if I can tell that it's AI generated, I'm much less likely to be interested in the product, whether it's a game, a video, an article, or whatever. As soon as that's clear to me, I click away. And I hope that enough people doing something similar sends a message to those publishing these things. But I expect that not too long from now I might not be able to tell the difference any more, and that because of economics, most of the bigger publishers will go that route anyway. I can continue playing indie games, but most of these will likely use AI art soon, perhaps because they can't afford a person either, or more likely, they would never even have started their game but for the essentially free art that AI provided them. I will continue watching videos on smaller channels that are clearly made by humans, but most of them will use more and more AI for the same reasons. Etc.
Crying won't make a difference. Passing laws likely won't make much difference either - there are always other jurisdictions. We're living in a world that is changing way faster than during the industrial revolution. What recommendations would you have made to people back then? Try to stay on top of things? Go with the flow? Use the new tech to your advantage? One thing is certain, not many people want to go back to living conditions before the industrial revolution. Given an equal amount of time, and assuming we survive until then, I doubt many people would want to go back to living conditions before the 2020s either. Maybe people losing their jobs is part of the price we pay for living in a better time.
PS: It's interesting that your disclaimer says "We do not *currently* use generative AI here at Molly Rocket." Do you see it as eventually inevitable as well?
It is unlikely we would ever use generative AI. It's impossible to predict the future, but if I had to guess, my assumption is that we will exit the industry permanently once (if?) generative AI becomes required. I have no interest in it, nor does Anna.
Since Demetri is a mathematician, it would be interesting to hear his perspective on math and AI.
Does AI affect the "Math for the Working Programmer" in any way?
[Long comment, feel no obligation to read it, though at least I can assure you it wasn't AI generated!]
It's incredibly easy to take a photograph these days. At least one that only a small handful of people would care about. It's much harder to take one that is in some sense attractive to a larger number of people. And it's incredibly difficult to take one that is beautiful, that takes your breath away, that causes you to study it for minutes or longer.
I can't think of anything AI could do to diminish the first of these types of photos. People taking selfies, or pictures of their families and friends, or even of their food - an AI image won't diminish these in any way, and would likely be frowned upon if it tried. That's not *me* standing in front of this particular structure or landscape, those aren't *my* friends, that's not the meal *I* ate. We might have AI generate them anyway, just for the novelty of making me and my friends look like superheroes, or making it look like me eating an 18" tall hamburger, but we'll always know they're not real.
I also don't think the last category, Ansel Adams photographs, for example, also won't be seriously diminished by AI generated images for the same essential reason: they might look pretty, be perfectly composed, etc., but they're not "real" in the sense that they don't reflect something someone saw in the real world. Here it's possible that AI might financially impact the livelihood of a photographer in that you could tell it to generate an image that resembles something like theirs, but I feel that the impact would be moderate at best. People will either know it's not real, not really taken by their favorite super-star photographer, or if it has been claimed that it was, it will ruin the reputation of the publisher when it inevitably is pointed out that it's an AI fake.
So that leaves the middle group. Let's call them professional photographers. Many of them probably take photos for a living. Photos that fill our books, magazines, catalogs, games, web pages, and backgrounds, for example. It was these various industries that created these jobs in the first place. These are the people that will be most impacted by AI generated images, and it's simply an economic thing - the people paying them would likely take *any* cost savings they can get, not simply AI that's now able to produce images that are similar enough that no one really cares. This has been the case since the industrial revolution started. I don't think there's anything that can be done for it because if one company/publisher for some reason doesn't join the AI bandwagon, while most everyone else does, eventually they will be left behind. Maybe they'll have some niche cache amongst a small group that insist their buggy whips be hand-made, but sales will be minimal, because in general, no one likes to pay more for something when they can't distinguish a difference in quality just because it saved some unidentified person's job. They have other expenses and possibly a diminishing income as well.
Other art will likely follow the same pattern. I'm not going to stop doodling on my coffee cups just because an AI can do it better. And super-star artists will continue to work in whatever medium they are known for, and there will be little impact in the price they can get for their painting/sculpture/movie, or whatever. The big problem here is that much of the middle-group's art is digital. (That portion that isn't, probably is safe for a while yet. At least until robots get a lot better.) Digital copies don't cost anything, whether they are literal bit-wise copies, or knock-off imitations in the style of some artist created by an AI. And again, it is the publishers that created the market for these sorts of art in the first place that will eventually be forced by economics to reduce what they pay for it. Except for niche cases, if they don't, they'll simply be priced out of business, and the artists will be out of work anyway. I feel sorry that these artists will have to find other ways to make a living. They work in a field where their product has been commoditized and better ways of producing it are now coming available. I similarly felt sorry for all the machinists that lost their jobs because industrial robots now do it better, or because their jobs were shifted to a cheaper country where, if not yet, soon robots will do it better. Soon I suspect I will feel sorry for all the doctors that will lose their jobs as AI starts doing a better job than they do. (I won't feel very sorry for the lawyers who lose their jobs, though =) And I feel sorry for all the people that work in the programming field because they were told it was a lucrative career. I'll be looking forward to hearing your take on this, and how it compares to artists' losses.
But all our feeling sorry for these people never brought any of their jobs back. We could try throwing our metaphorical wooden sabots into the metaphorical gears of AI data centers, but it's not going to make a difference long term. Personally, if I can tell that it's AI generated, I'm much less likely to be interested in the product, whether it's a game, a video, an article, or whatever. As soon as that's clear to me, I click away. And I hope that enough people doing something similar sends a message to those publishing these things. But I expect that not too long from now I might not be able to tell the difference any more, and that because of economics, most of the bigger publishers will go that route anyway. I can continue playing indie games, but most of these will likely use AI art soon, perhaps because they can't afford a person either, or more likely, they would never even have started their game but for the essentially free art that AI provided them. I will continue watching videos on smaller channels that are clearly made by humans, but most of them will use more and more AI for the same reasons. Etc.
Crying won't make a difference. Passing laws likely won't make much difference either - there are always other jurisdictions. We're living in a world that is changing way faster than during the industrial revolution. What recommendations would you have made to people back then? Try to stay on top of things? Go with the flow? Use the new tech to your advantage? One thing is certain, not many people want to go back to living conditions before the industrial revolution. Given an equal amount of time, and assuming we survive until then, I doubt many people would want to go back to living conditions before the 2020s either. Maybe people losing their jobs is part of the price we pay for living in a better time.
PS: It's interesting that your disclaimer says "We do not *currently* use generative AI here at Molly Rocket." Do you see it as eventually inevitable as well?
It is unlikely we would ever use generative AI. It's impossible to predict the future, but if I had to guess, my assumption is that we will exit the industry permanently once (if?) generative AI becomes required. I have no interest in it, nor does Anna.
- Casey