Things are definitely changing a bit faster than I expected. And, it is easiest to see these changes if I look as far outside my typical day-to-day tech usage as I can. Once I do, I find things like this Tom Scott video…
I am a bit older than Tom. And, I do remember everything that he mentions. And, I also remember the process he describes happening with earlier technologies (Video Cassettes, CDs, etc). And, I agree with Tom. The world is about to change.
“All it [ChatGPT] does is predict what the next word is going to be.” — Tom Scott.
And, like Tom, I tried ChatGPT on some code. And, I too found that it made some mistakes, not horrible ones. I explained the error to ChatGPT and it mostly fixed my code about as well as it fixed his.
“I don’t think I’d have a problem with my work being a small part of a massive pool of training data. But if someone started ripping off specifically my name and my work? Yeah, I can see why artists are furious about that.” — Tom Scott
I am a bit more hesitant than Tom about the first part of this statement. When I created and shared data before the advent of modern AI, I had expectations about how that might be consumed. In fact, that is why I chose the Creative Commons license that you can see linked at the bottom of the site footer. And, I feel that if the companies are building systems that sample any work licensed as mine is or explicitly under copyright, etc, then the companies making money through what is derived from that data being part of their model need to pay for use of that Intellectual Property.
“[T]hat feeling of dread came from the idea that ChatGPT, and the new AI art systems, might be to my world what Napster was to the late nineties. The herald, the first big warning that this new technology, the thing that was going to change everything.” — Tom Scott
I think Tom has distilled the essence of what seems to be worrying many working artists today.
Are these tools for most of us? Or, will these tools replace most of us?
The metaphor that I tend to use when discussing this goes back much further than the 1990s. In the earliest days of cinema (the movies), the most common form of entertainment was Vaudeville.
The biggest problems these theater owners faced all concerned the talent they hired.
And, if a slick salesman showed up with a movie projector and said to that owner, “You get to pre-screen the film to be sure the content meets your standards of quality. You have a guarantee that the exact same show will be presented for performance after performance. And, there will never be a missed performance again! Finally, instead of paying 20-30 people, you only need to hire 1 projectionist (and before talkies a pianist to accompany the films).”
Movies were cheaper, more reliable, and demonstrably more consistent than Vaudeville ever could be.
So, some early adopters changed over their theaters and got flack from angry performers. And, over time, the technology improved (talkies, color, etc).
And, if you can name any Vaudeville performers off the top of your head, they are likely within the group of performers who figured out how to follow the work to another medium (movies, radio, TV, etc).
So, where are we on that Sigmoid Curve that Tom Scott mentions?
If we are close to the end, then he’s right, these are tools that will improve the lives of most artists making a living through their art today.
But, if we are in the middle of that curve, then Tom is likely also right and the best advice to give to working artists is to figure out how to adapt what you do to whatever new medium starts to arise.
To be clear, AI isn’t replacing artists. AI is likely to get hired instead of artists to do many of the things artists now do to make a living from their art. Art is not dying. The compensation model for art is what is likely to be crushed as AI picks up speed and barrels along that Sigmoid Curve.