I don’t exactly understand your question, but I’m assuming you mean something like, “What value is this person actually creating?”
I can think of two direct use cases where I would find it helpful:
- Creating a “précis” (concept from Neuromancer) video or article on a particular topic. For example, let’s say I want to learn about pre-20th century Indonesian art. Sure, I could Google and read some articles, but it would be vastly more enjoyable and efficient to just get a 5-minute video auto-generated on the topic.
- For creating content that I want to read but don’t necessarily have the time to write. I’d love to create a website that catalogs subcultures, for example, but I just don’t have the time to do the research and writing myself.
Both of these are dependent on the generated content being accurate, of course, but that seems like a temporary problem, not a flaw in the system itself.
I think the nearby comment that points out that this is more closely related to editing than to creation as it has recently been understood is correct. Which is not a value judgement per se, but the distinction between those two activities is useful and to me provides a useful lens to understand this change.
On the face of it, that is true. But on a deeper level, AI generation is really just a kind of rapid sketching, and sketching is a practice that goes back millennia.
Additionally, it depends on how you are defining "creation" and there have been various artistic movements which "created" art in a manner not dissimilar to AI image generators. The cut-up technique or some Dadaists, for example. I certainly would consider works by Burroughs or Tzara to be real creations and not simply editing.
You're conflating two meanings of sketching though. Artists and others do uses it as an exploratory tool, and I've already seen AI generation used in a similar way so I know that's a valid application of it.
But at a lower level sketching is itself a visual interpretive and creative process. And yes what we're talking about is mechanically similar to the cut-up technique, but is that used for "sketching" in the exploratory sense?
I don't know, maybe. My point overall is that these tools aren't "really just" anything. They have similarities and similar applications to many things we've seen and done, yes. But they are fundamentally alien and in important ways are unlike anything else we've done. We should be careful about and aware of the limitations we put on our understanding when we try to interpret them through the lens of what has come before. To some extent we have no other lens, but that doesn't lessen the limitation.
I think in this case the role of the human in the process does have a strong analog to that of an editor. Where you have the ability to initiate and then shape the creation of another entity, but the creation itself is out of your hands. It's not exact, but it may be close enough to be useful, in a way I think comparing it to the cut-up technique is not quite.
I can think of two direct use cases where I would find it helpful:
- Creating a “précis” (concept from Neuromancer) video or article on a particular topic. For example, let’s say I want to learn about pre-20th century Indonesian art. Sure, I could Google and read some articles, but it would be vastly more enjoyable and efficient to just get a 5-minute video auto-generated on the topic.
- For creating content that I want to read but don’t necessarily have the time to write. I’d love to create a website that catalogs subcultures, for example, but I just don’t have the time to do the research and writing myself.
Both of these are dependent on the generated content being accurate, of course, but that seems like a temporary problem, not a flaw in the system itself.