This week, Anthropic announced that Claude will now include invisible watermarks to A.I. text and images. This news has been getting around, and it comes off the back of the A.I. music generator platform, Suno, pretty much saying the same thing.
Suno’s PR team has framed the watermarking as an effort to combat slop. But we all know the real reason is because of legal issues. The company recently lost a landmark copyright case last week, one which has a high probability of setting a precedent on the ‘fair use’ argument around A.I. ‘training’. (Hint: none of these A.I. companies have been right about fair use, and they’ve known it, hoping to just become so ubiquitous that they would get away with it—a typical move from the tech industry.) Not to mention that Suno also settled with Warner on a lawsuit which probably included some idea that Suno-made songs should be watermarked (plus, they’re limiting downloads now, upsetting the apple cart for heavy slop makers).
Today, Spotify announced it will not recommend music from A.I. personas, and will put an “A.I. persona” badge on these sorts of artists.
Last month, Substack implemented it’s built-in A.I. detector. And just about every streaming service has now opted to label A.I. in some way. Not to mention that Bandcamp were an early mover and banned it outright.
The market has spoken, and I think the platforms have realized what many of us have been saying all along: generative A.I. will only drive up the value of human-made work. The market prefers human-made, or at the very least, feels like it should know when something involves A.I., which is a perfectly reasonable request.
A friend of mine messaged me this morning and said these latest moves reminded him of the opening scene in Blade Runner. The Voight-Kampff test is one used by the LAPD Blade Runners to help determine if a person is bio-engineered—a “replica”. We’re really doing the same thing with A.I. now.
One of my friend’s points about this, however, is a lot of this is based on fear, and fear often leads to the exact outcome you’re trying to avoid!
Whether these moves are based purely on fear or just on market dynamics, I don’t know. It’s probably a bit of both, because fear can really drive the market.
Since the platforms began to move toward labeling A.I. content, I’ve seen a lot of complaints from people who use the technology. These complaints tend to ask the question: what’s the difference if I use A.I. or not? For example, in writing we use editors and often whole teams to make writing better. What’s the difference if you use A.I. to do it?
Last year, I laid out some of the ethical challenges to using A.I. when an A.I. tool has scraped everyone’s content without permission. The idea that there is no difference between using A.I. or having other humans involved in the creation of something is false. The difference is in the very premise—you are using A.I., not a human. That’s a difference.
The tech industry is notorious for trying to convince us that when tech mimics what human beings do it is ‘the same’ thing as a human being doing it. But it’s not, because it is a technology, not a human being. It helps to call something what it is. Napster was never ‘just like’ giving your friends a mix tape, and social media is not ‘just like’ seeing your friends in real life. That doesn’t mean these things aren’t valuable, only that they’re not the same.
There is a big difference if I tell people the editor on my latest book is President Donald Trump or the editor is my mom. Yes, they are both editors and the job would have been fundamentally the same, but there is still a difference. Whether or not the difference is one of quality is one only the reader can ultimately decide (my mom will do the better job, by the way!) but it’s only fair that people want to know. It informs them on whether or not they want to invest their time, money, effort, or not. A good editor helps my reputation as much as a bad editor may harm it.
Likewise, whether or not you used A.I. informs people. It tells them something. For many, that means they don’t want to read your work. There is no sense in complaining about that. It’s the same as if you make jazz music and complain that Swifties don’t like it. You’re not making music Swifties will like, but you are making music jazz aficionados might like.
What about people with health issues, or disabilities, or financial issues who can’t afford a human being to help them create? Should they be punished for using A.I.? The answer is no, and being open about it does not mean they will. What it means is the reason and intention behind the A.I. use can be made more clear and that reason can often work to your advantage. Because, art often sells not just because of the aesthetics, but the story behind it.
That’s why Neil Harbisson, the world’s first recognized Cyborg, gets respect for his cyborg art—not because solid palettes on large canvases are necessarily in demand aesthetically, but because he was born with a disability that prevents him from seeing color, and his antenna helps him ‘hear’ sound, which he translates onto canvas. More at this video below, but I’ll probably write a larger article on this (I am writing about him in my upcoming book on culture, art, and pastoring).
Again, Donald Trump will ask me for money to edit my book, but my mom might not (or she might)! I would probably not be punished for having my mom edit the book. In fact, in some cases, I might even be rewarded for it, depending on the final outcome, because many will think it was sweet that my mom edited the book and that’s one thing that will make them connect to it.
In the long run, I suspect that those who need to use A.I. not because they’re trying to get thoughtless content out quicker and game the algorithm, but because of human fragility, may even be rewarded for how effectively they use the tech.
A.I. has not devalued human-created art. It has done the opposite, and it will continue to do so. Art’s value will always be judged on the human element—the human story behind the art. The silver lining with A.I., and I’ve also been saying this from the beginning, is that it’s going to finally bring the content slop algorithm machine introduced by social media to reckon with itself. The age of the online social media slop-fluencer is coming to an end, and A.I. has helped us see through it. Good riddance, I say. The “new media” revolution will come full circle.
For more thoughts on this, check out this essay:
Comments
Nothing yet. Say the first thing.
Sign in to join the conversation.