Daily Pulse

The Year In AI: The Revolution Will Be Digitized As Debates Ethical, Legal & Artistic Swirl

y.robot
The Robot Band performs for the audience at World Robot 2025 in Beijing, China, on August 9, 2025. (Photo by Costfoto/NurPhoto via Getty Images)

Trying to keep up with all the developments in artificial intelligence in 2025 is a task analogous to trying to grab the tailfin of a rocket with a lacrosse stick from the back of a pickup truck. We can see how high and fast it’s going, but we aren’t yet armed with the tools we need to grasp it all, let alone keep up.

Still, 2025 is destined to be remembered as a watershed year for the rapidly proliferating technology, even if, with the benefit of hindsight, it turns out to be just one in a series of watershed years.

Debates about whether AI-generated music is art in a meaningful way are probably best left for philosophers and academics and may not be answered to anyone’s satisfaction in any substantive manner for years, but the more prosaic concerns of commerce and legalities are already being tackled and in the nick of time. Social media — on both the AI doomer and booster sides of the debate — breathlessly reported that AI had its first chart-topper: “Walk My Walk” by Breaking Rust (both song title and artist name come off like parodies of what AI would come up with, except they managed to be spelled correctly). And that’s technically true. The song did top the Billboard Country Digital Song Sales Chart. That’s an alarmingly specific chart, and all praise to our counterparts at Billboard for tracking things like “country digital singles sales,” but, of course, that’s not the country chart (which, as ever, was topped by Morgan Wallen) and digital singles sales are a teeny, tiny part of the market. The song was nowhere to be found on the main chart nor on Spotify’s streaming charts. Now, Xania Monet — another AI creation — did debut on the R&B radio chart, coming in at No. 30 in early November.

Maybe the revolution is coming, but it’s not quite at the doors of the castle yet. Still, it was high time to grapple with all the monetary and legal ramifications of the technology.

A foundational question remains if people’s likenesses — and ultimately their voices — can be used by generative AI without explicit permission. OpenAI — the entity behind ChatGPT and video generator Sora — and most of its competitors prefer an opt-out approach, where artists’ likeness would be included by default and removed only by explicit request, while the entertainment industry would prefer an opt-in regime. The EU, which is ahead of its American and Asian regulatory counterparts on the issue, seems to be leaning toward opt-out. WME wants no part of it, telling OpenAI to remove all of its clients, in all sectors, from Sora. 

Meanwhile, the three major North American performance-rights organizations — BMI and ASCAP in the U.S. and SOCAN in Canada — announced they’d accept registration for “partially AI-generated songs,” while rejecting any wholly created by AI, reflecting the broader policy from the U.S. Copyright Office, which said wholly AI creations are not subject to copyright protection. Near year’s end, the three majors — Sony, UMG and Warner — finalized a licensing deal with L.A.-based Klay for its music platform that will “help further evolve music experiences for fans, leveraging the potential of AI, while fully respecting the rights of artists, songwriters, and rightsholders.”

All this wrangling comes as French streamer Deezer — which has an industry-leading AI detection model as it seeks to tamp down on the proliferation of AI-created songs to more fairly calculate its royalty-payment pool — said nearly a third of new music uploaded to the platform is AI-generated and that a multinational survey found that 97% of listeners couldn’t distinguish between human- and AI-created music.

Which raises a disturbing question: when the first AI artist actually tops an actual chart, will anybody be able to tell?

FREE Daily Pulse Subscribe