“I’ve heard artists say this. ‘I don’t use any AI. I don’t use any Suno for my stuff.’ Bull crap. I’m telling you, I’ve seen singers out there give interviews where they’re like ‘I don’t do that. I don’t do that.’ I’m telling you for a fact, LIE. They’re lying through their teeth. Maybe at some point they said, ‘I’m only considering THIS to be AI or Suno or whatever. But yeah, they do that. Whether they know it or not is sometimes another story. But they do.”
These were the recent revelations from country artist Joe Nichols talking to Evan Paul at Taste of Country where the country traditionalist also revealed how Suno played a role in his new song “Say La V.”
According to Nichols, “Say La V” writers Brett Warren, Jason Sellers, and Lance Miller did not use AI to write the song itself, but did put it into Suno to spit out a demo they could then send to Nichols and other prospective performers while pitching them to record it.
“They pulled the song together as a joke,” Nichols explains. “Then they came up with a catchy melody. And then they put it through one of those AI-generating, ‘What would this song sound like if it had these prompts?’ And it came out really super catchy.”
They then sent the Suno demo to Nichols who at the time was the chaperone on a Washington D.C. trip with one of his kids, but listened to it, and sent it back saying, “Too pop!” But he kept humming the melody, and as time went on, he had a change of heart and decided to record and release a more “country” version of the song. “Say La V” came on on July 17th. And frankly, it still sounds pretty pop.
The revelations in the interview have some worried just how much AI assistance is creeping into the songwriting and production of country music, especially if a more traditional and older artist such as Joe Nichols is working with it, let alone the 20-somethings looking to score their first hit. Nichols goes on to reveal,
“I’ve written some things that I put into Suno or one of those to see if I’m looking at it the way other people would look at it, the algorithm would look at it. Do I hear this song as a bluesy Gospel song, or does it come out as a bluegrass song, and a different vibe completely? So it is helpful to tell you if you’re on the right track.”
Nichols then clarifies, “I don’t go any further with that with Suno. I don’t want it building anything for me. I want it to tell me where my blind spots are, what I’m not paying attention to … the singers on Suno are amazing. I don’t use those because the singers from Suno are the best singers in the world combined as one singer. So it’s impossible to do something as good as Suno.”
But there are still some moral dilemmas when dealing with Suno in the music creation process, even if it’s not being used to create the finished song itself. First, most all record labels—mainstream and independent—have pending lawsuits against Suno for ripping off their copyrights, and the copyrights of artists like Joe Nichols. Suno uses the existing catalog of recorded music to compose its prompted songs in a matter of seconds, and does so royalty free. So as record labels sue Suno, the artists signed to them are using and supporting the tool.
The second dilemma is if a song was written by humans, but used a Suno demo to either flesh out the writing or influence the production, does this need to be disclosed as “AI-assisted” under the recently universally proposed AI labeling, even if Suno is not writing or populating the music of the song in full?
Also, along with not considering the slippery slope of people who rely on copyrights to protect their intellectual property like songwriters and performers, there are other downstream effects to consider. The building of data centers has become an almost universally-repudiated practice on a local level. Those data centers are being built in part to build out the infrastructure to support songwriters using Suno to create demos, and deliver them promptly.
On Tuesday (7-20), streaming service Deezer—which has been in the lead detecting and tracking AI-generated content—said that for the first time, the amount of AI songs uploaded to the format was over 50%. Even though many of these songs never go on to be heard or become popular, it’s still a massive volume, equating to about 91,000 tracks per day. The company also reveals that often, the majority of streams these tracks receive are fraudulent, up to 85%.
Later in the conversation with Joe Nichols, Evan Paul of Tastes of Country asks him, “But what’s the difference between Auto-tune though, and singing it and having it Suno your voice?”
Joe Nichols responds, “Ultimately, I don’t view it much differently because Auto-tune saves time. But here’s the thing about Auto-tune that really drives me crazy. We have developed a generation of singers that sing like Auto-Tune. The nasally pitch and pinch in their vocal.”
Auto-tune was never originally meant to pitch correct an entire performance, either live or in the studio. The point of it was when you had a singer that turned in an incredible performance, but perhaps a note or two was flat. Auto-tune allows you to save the performance, but correct the pitch on those particular notes.
Some have equated Auto-Tune to AI, but it’s a completely different tool. Auto-Tune can’t create a vocal performance for an entire song in seconds like Suno, it can only correct the pitch of an existing one. But similar to how some artists might sing to mimic Auto-tune, the use of AI could also influence how human creators write and perform songs, even if they don’t use the technology directly.
Joe Nichols continues on to say, “Ultimately, if you’re using electronic help, you’re using something else that’s not organic art. But to me, something that’s AI—something that’s not real, people hear. It’s inauthentic. This is why live music is our last frontier. And it drives me insane when I see other artists on these festivals that I play on, and I hear them out front and they sound amazing. I’ve never heard this person sing like this in years. They sound so good. And their crew guys are like, ‘Yeah, of course we’re ‘tuning it.’”
Nichols continues, “[Live is] the one time that you can be honest and authentic, this does not save time. This just makes you look like you’re perfect, and that’s bull crap. If you’re not good enough to come out here and do it 90 minutes, then you’re not good enough. Get better. I can’t stand that. I don’t do that, and there’s a lot of artist that don’t. But there’s a lot of artists that do.”
But Auto-Tune started as a time saver in the studio. Then it became a crutch for certain performers who can’t sing to create an entire career from. A similar slippery slope is presenting itself for AI. Today, songwriters can use it to build out demos, or test songs for appeal “in the algorithm.” Tomorrow, it could be full integration with the software to replace the need for co-writers, or musicians, and ultimately, any human as AI reveals it can do everything in the music making process, and not complain about copyrights or coffee breaks as music creeps toward full digital integration, leaving the human element that’s so integral to music behind.
– – – – – – –
If you found this article valuable, consider leaving Saving Country Music A TIP.
‘ The preceding article may include information circulated by third parties ’
‘ Some details of this article were extracted from the following source savingcountrymusic.com ’
