Interesting article by Alberto Romero on how visual artists and writers seem to be reacting differently to ChatGPT and like technologies. In particular, writers seems less threatened by it.
Generative AI is expected to impact all kinds of jobs. Artists, writers, and coders may be in the danger zone. While some workers see progress and productivity, others resent technological progress as it threatens their way of living.
Romero discusses the theory that visual, artistic styles are easier to fake, at least to amateurs, than writing styles (fake replicas may in fact be obvious to trained artist’s eyes). Like a field of bumper cars, there are many drawing styles and choices an artist might make. There is much freedom. Writing, the theory goes, is like railroad cars forced to move down a single track in a well-ordered fashion.
You have to go down that exact path for the writing to make sense. Any small deviation from proper language is almost certain to result in a nonsensical statement. There is more flexability with visual art.
Thus, artists may be particularly vulnerable to AI’s outstanding ability to remix data and create something sufficiently original that still avoids plagiarism, while not losing the feeling of a familiar style. While generative AI can copy styles superficially, artists fear that non-artists wouldn’t care about that superficiality.
In contrast, writers are safer from generative AI because text doesn’t rely on style as much as images.
I can imagine another reason why writers might feel less threatened. ChatGPT is more of a tool for writers than for artists.
It is easy to see how language technologies can be used to assist writers by scouring the Internet for relevant information and summarizing it, taking raw items or outlines written by humans and providing fuller coverage or even by suggesting potential extensions or new ideas to existing themes. This is a rather straightforward use of the technology. Deploying the same kind of technology for pictures is more difficult.