A while back I saw a video of some Japanese animators trying to do a similar experiment, and both of these experiments have gotten me thinking.
When AI can create characters that are human enough, when they can render scenes that are realistic enough, and they can stumble into interactions between these objects that are interesting enough, will they completely replace human artists? Will entertainment basically become just watching an AI play a system like Minecraft (with much higher fidelity)?
On the one hand, it’s an exciting idea. Set up the right systems with the right kinds of actions and interactions (or heck, let an AI figure that out too, why not), and let the AI loose to explore all the different possible things that can happen. Create another AI that can recognize which of the possible set of actions would be interesting to an audience, and boom, you have an infinite content machine. You could do things like create a model of Middle Earth and have millions of stories about dwarves and hobbits (but not elves; they’d be removed by the interest recognition system).
On the other hand, it would rob the humans of the joy that comes from creating media. Like it or not, anything we can train a machine to do, it will do better than us. This is the whole reason we made them. I don’t think that the automation of media would effect the “high arts,” because the idea of high art is based entirely upon a social construct*. The people who would suffer would be the people who do the media jobs that effect us all. It would harm animators, film makers, and game creators. It could very easily be a blow to illustrators, pop musicians, and novelists as well. These jobs require audience attention and retention in order to function, and if there are infinite content machines that can be tuned to create content that is massively appealing (or–even more dangerous–content that is specifically appealing to every individual), I find it very hard believe that human creators will be able to keep up.
A way to work against this potential outcome is that audiences come to value the people that are creating things as much as they value the thing being created. This may inoculate the pop music industry, for example, because pop music isn’t about artistic appeal, but human interest. The obvious problem with this, is that when value of media is determined only to the person performing it, you’ll end up with objectively simplistic media. Once again, see pop music. The media being created isn’t the product, the person is. The media is just an excuse for the audience to throw money at the one sided relationship with an idol.
Another thing that may happen is that human made media will continue to be consumed by niche audiences that follow it specifically for the meta-content that it contains. The story of how a piece came to be in the context of the culture and media tradition are what they’d really be consuming. I also find this troubling, however. I prefer as little context as possible when consuming content, because I like to let content speak for itself. I want to bring my own context, and have content speak to me on its own without being bogged down by meta analysis.
Anyway, it’s reality, and it’s problematic. Media creation has already been made crazy by the destruction of local markets, why not add some more craziness in by introducing robot creators as well?
* The AIs wouldn’t be able to prevent people from wanting to show their differences in class based upon their art consumption.