In recent months, we’ve been bombarded with millions of lousy AI songs, idiotic AI videos, and clumsy AI images. Error-filled AI texts are everywhere—from your workplace memos to the books sold on Amazon.com. (...)
 |
| All Fake |
The garbage hasn’t disappeared. It’s still everywhere, stinking up the joint.
But people are disgusted, and finally pushing back. And they are doing so with such fervor that even the biggest AI companies are now getting nervous and pulling back.
Just consider this surprising headline:
This was stunning news. YouTube is part of the world’s largest AI slop promoter—namely the Google/Alphabet empire. How can they possibly abandon AI garbage? Their bosses are the biggest slopmasters of them all.
After this shocking news reverberated through the creative economy, YouTube started to backtrack. They said that they would not punish every AI video—some can still be monetized.
But even the revised guidelines are still a major blow to AI slop purveyors. YouTube made clear that “creators are required to disclose when their realistic content is altered or synthetic.” That’s a huge win—we finally have a requirement for disclosure, and it came straight from the dark planet Alphabet. [ed. who's motto used to don't be evil]
YouTube also stressed that it opposes “content that is mass-produced or repetitive, which is content viewers often consider spam.” This is just a step away from blocking slop.
What happened?
Maybe the folks at YouTube are just as disgusted by AI as the rest of us. Or maybe we have shamed them into taking action.
My view is that YouTube is (finally) reading the room. I’ve noted before that YouTube is the only part of the Google empire that actually understands creators and audiences. And (unlike their corporate overseers) they have figured out that AI slop is an embarrassment that will tarnish their brand.
The widespread mockery of the fake AI band Velvet Sundown might have been the turning point. This blew up in the last few days, and left AI promoters reeling.
Velvet Sundown is a non-existent AI band that got a million plays on Spotify. These deceptions have occurred in the past, but something different happened this time.
Music fans started mocking Spotify and its alleged promotion of a stupid slop band. The company was subjected to a level of ridicule and angry denunciation it has never endured before.
Journalists called this out as a hoax or fraud. And many speculated about Spotify’s role in the charade. After all, the company has been caught promoting AI slop in the past.
But this time Spotify got turned into a joke—or even worse. They were linked to a scam so clumsy that everyone was now making fun of them, as well as scrutinizing their policies and practices.
Rick Beato’s response to Velvet Sundown got two million views—so more people were watching takedowns of the band than listening to it. An industry group even demanded disclaimers and regulation.
And the jokes kept coming. People mocked the slop with more slop
That must be painful to endure, even for the billionaire CEO of a streaming platform.
Whatever the reason, Spotify started to buckle. It actually began imposing restrictions on AI.
“Spotify has now pulled several uploads from the AI act and the associated Velvet Sundown,” reported Digital Music News on July 14.
It felt like the tide was now turning in the war against slop AI music.
Dylan Smith, one of the best sources on this subject, clearly thinks so. “Velvet Sundown’s Spotify pulldown,” he writes, “doesn’t exactly bode well for forthcoming AI releases.”
I’m focused here on AI’s destructive impact on culture, but there are other signs that growing AI resistance is now forcing companies to reconsider their bot mania.
“An IBM survey of 2,000 chief executives found three out of four AI projects failed to show a return on investment, a remarkably high failure rate,” reports Andrew Orlowski. “AI agents fail to complete the job successfully about 65 to 70 percent of the time, says a study by Carnegie Mellon University and Salesforce.”
He also shared the results of a devastating test that debunked AI’s status in its favorite field, namely writing code. This study reveals that software developers think they are operating 20% faster with AI, but they’re actually running 19% slower.
Some companies are bringing back human workers because AI can’t deliver positive results. Even AI researchers are now expressing skepticism. And only 30% of AI project leaders can say that their CEOs are happy with AI results.
This is called failure. There’s no other name for it.
And it will get worse. The Gartner Group is now predicting that 40% of AI agent programs will be cancelled before 2027—due to “rising costs, unclear business value and inadequate risk controls.”
by Ted Gioia, The Honest Broker | Read more:
Images: Bridge Chronicle/YouTube
[ed. I'd say temporary setback. The AI industry will eventually figure something out, they've got too much money and tech beavers involved not to. The product will get better, legislators will be lushly rewarded for IP protection and distribution, some hit movie/song will get made entirely by AI, some important (maybe unusual) event will occur and eventually be traced to it, etc. A million things could happen. So calling this winning seems a little premature. Likely we'll just get used to it over time (like advertising), with authenticity mostly a certification issue (if anyone cares. you have to wonder with taste these days). See also: I'm Sorry... This New Artist Completely Sucks ie. how to create a fake song of your own (with just two sentences) (Beato)]
