But at least photography created new roles for photographers. AI art promises to replace many creative roles for humans entirely, without creating any new roles.
These AI art tools model distributions based on sample data. They don’t create the underlying distributions; they need to be given (many) samples from the source distributions.
You need people even just to say if something is good or not. Standards will rise considerably as AI tools mature, and peoples’ skills with the new tools grow. Things get old real fast and people get bored quickly.
Here’s a new role: creating trademarked artistic styles. People will become tired of what eventually becomes seen as a limited set of “AI styles,” which are actually distilled from the techniques of a relative handful of human artists. So they will seek out “Natural” artists who are able to create new artistic styles. Also, the Disneys and the mega-studios will need and want to create distinctive and commercially protectable visual styles.
The legal framework and case law will have to be worked out. Think of it as “Look and Feel” on steroids. One thing’s for sure, it will create significant roles for creative IP lawyers.
The examples you gave sound like they would only be for the top tier of today's artists though, or a small number of AI taste supervisors, not the general workforce. A company would need to hire far fewer people if all they need to do is feed a new style into the image generator occasionally. I also don't see that as a new role created by this technology, since people are already doing that - those are just a few people who might not get laid off.
If standards rise considerably it could help, but I'm not really able to picture what that would look like, in a way that would mean more roles for people. Or if the tech gets stuck where it is now and never becomes truly useful, then I could see AI as just being another tool.
The invention of recording devices gave rise to the recording artist. Only a relative handful of people can be super successful at this, but there is a greater population of working musicians and singers.
YouTube and TikTok, smartphones allow millions of people to become influencers. Some people make millions playing video games live online on sites like Twitch. Only a handful become rich through new media, but it seems to allow a fair number of people to make a living. Certainly there are far more minor celebrities than ever before. Even (really talented and/or self-promoting) chefs are celebrities now.
I’m sure people will take classes, hire instructors / tutors, buy books, watch YouTube videos on how to get better and/or more efficient with AI art and design tools. It’s not enough to teach people new technical skills. You also need to teach them how to monetize those skills.
The key takeaway is that the barrier to entry has dropped through the floor. There will be an explosion in visual content. In a world where nearly everyone is at least an amateur content creator, and probably at least internally a content creator at work, there’s going to be a skills pyramid and strong demand for legitimate talent.
Well, I agree that some technologies do justify themselves by creating more jobs/roles than they take (I think you're overstating the number of people making a living as social media streamers/influencers though). But the question is whether AI art specifically is going to do that, or if it will simply remove jobs and not add any back. People might still buy books and train themselves, but if the demand for art can't scale up, they won't be able to find good jobs or creative fulfillment.
I didn’t state a number of people making a living on social media. Some people are clearly making a fortune; how many people who are making a living off it, I don’t know. Enough so that I typically have no clue who some random Instagram, TikTok, or YouTuber with a million followers is.
Historically, new technologies create more jobs than they eliminate. Books used to be copied by hand, by the relatively few people who could read and write. The printing press eliminated a lot of the bread and butter scribe work.
Employers love finding ways to eliminate labor. Autonomous vehicles promise to eliminate millions of driving jobs. But eliminating labor is not an end, it’s a means to an end, which is increasing profitability - a function of productivity.
Employees have stayed ahead largely by boosting their productivity faster than labor saving technologies eliminated jobs. The same technologies which eliminate jobs tend to simultaneously increase the productivity of other jobs. Employment is at an all time high.
I have no idea what all the new jobs created around AI art and design tools will be. I presume it will create a new subfield of law. There will be new laws and regulations, regulators, some compliance burden on businesses (e.g., are they illegally exploiting unlicensed commercial content), auditors. There will probably be reviewers and editors for AI generated content. There will probably be considerable AI infrastructure, consultants, support personnel, administrators, AI model and data providers, annotation services.
For most office workers, the level of graphics and visual elements in reports will probably go way up. PowerPoint will be a distant memory, but expectations regarding presentation quality will seem insane by current standards.
> I didn’t state a number of people making a living on social media.
Sorry, I misread your implication then. Twitch had a leak a while back and something like 4500 people were making over $50k USD, which is not nothing, but also not that practical as a career strategy.
Hopefully you're right that AI will open lots of opportunities for everybody, and hopefully those opportunities will be worthwhile. I know historically technology hasn't brought about the end of all jobs, but AI is a pretty new concept and we can't necessarily scale up forever. I think it's reasonable to be concerned about the number and quality of jobs remaining. It does seem like some of the jobs that people really want to do, are some of the ones AI could be taking.
But I don't really know one way or another, just feeling concerned and a bit bummed out about it. I guess we will see over the next 5-10 years.