music – The Human Project

Part 11: Lawsuits Against Suno, Potential Outcomes, and Where AI-Assisted Music Goes From There

We’re at a fever-level crossroads in the way of AI, and many of the bigger issues on how to regulate it all (you know, about it potentially destroying humanity) are still up in the air. In the realm of music, however, it seems some general consensus is being met, despite some lawsuits still pending, particularly at Suno. As the leading source of AI-assisted music creation, Suno has a major tug-of-war going on with some industry support—while others are still going after it for absorbing prior voices/songs like a digital sponge.

The most interesting thing happening with Suno now is it has one power player working alongside them in properly licensing music. On the other, you have another major corporate entity still going after Suno for using their earlier training method of accumulating existing musical IP. Chances are looking better that recording labels may fall in line with Suno over time based on a new business structure of licensing existing recordings.

Much of this can only happen because Suno is exploding in subscribers and general users—which includes myself to create studio-quality recordings of my own songs.

Regardless, the question remains of whether Suno can still manage to overcome all their other lawsuits—not including the growing public aversion to AI-created songs.

The New Warner Music Deal

If you use Suno in any capacity, you know they just released their V6 model. While some attest V5.5 was better, that old version is now retired. And it’s a symbolic gesture since V6 brings a new way to use prior existing music to the Suno app. They just made a deal with Warner Music Group, BMG, and Believe to license all their music for use in AI creations. The deal ensures the original creators of that music get compensated for the collective use on Suno for AI-created music.

Perhaps that seems like a major capitulation on the part of several major music labels. Regardless, it was always about money over the ethical qualities of creating things with AI. You can debate forever about whether it’s still an ethical practice, despite making it all about everyone financially benefiting on both ends.

It’s this development that now retires the old system of Suno’s training model using music without proper licensing. Even I admit the studio recordings I made of my first two songs seem like a compilation of existing voices, instrumentations, and choir sounds. I even went so far as to say the male singer on my single of “Whispered Thoughts” sounded like—um, Drake Shelton.

Some critics now say the V6 update is less superior to V5.5, though. I haven’t tried V6 yet on my prior songs or any new ones. A few users say the quality is more basic over the more believable voices and studio sounds 5.5 had. There’s a possibility that Suno had to lessen the quality just to stay alive. Prior to planning this segment of The Human Project, it appeared that Suno was soon going under with mounting lawsuits across the board.

This isn’t to say they still don’t have legal hurdles to overcome. They can’t completely erase their earlier model system that probably makes up over a billion or more songs already on streaming services around the world.

Other Lawsuits Still Pending

Last year, Sony/Universal filed a major lawsuit against Suno based on latter using copyrighted material in their original AI training methods. Suno can’t immediately shake this lawsuit based on all the songs already created using that process. It includes those who merely use Suno as a technological assist (like me) to make their written songs sound like they were recorded professionally.

In a time when the expenses for studio time are beyond many budgets, being able to hear your songs fully realized on Suno has been a major creative convenience. Not that I didn’t realize the sounds you hear on my two songs are a likely amalgam of prior voices and instruments. Sure, there’s a sense of ethical guilt about that, even if all future V6 songs may change everyone’s outlook.

Suno still contends that fair use was at play on their old training model, something you can argue as fact when mixing existing IP ultimately turns into something else. I waver in my legal beliefs on this since there’s still a major difference between creating voices and sounds from scratch rather than borrowing from them. But, with my own songs, I managed to create original songs that still sound original thanks to using that library of audio.

Whether Suno can beat this lawsuit in coming months (or years) is still up in the air. Their biggest pain is also being sued by performance rights organizations, particularly overseas. Those contend that Suno, through their prior model, didn’t properly compensate artists based on the audio taken.

Most notable there are lawsuits from performance rights organizations overseas and in Canada. For the latter, it’s SOCAN (The Society of Composers, Authors and Music Publishers of Canada) going after SUNO recently for reproducing and/or copying songs from the org’s catalog. They say over 150 works were essentially stolen as part of Suno’s training model. Then you have GEMA from Germany that went after Suno for similar reasons. In that case, Suno lost the case in saying they infringed copyrights.

Outside of America, Suno still faces some serious threats, even if the chances are greater they’ll overcome now with the above deals and new business philosophy. Ultimately, though, it’s up to the world populace on whether they really survive far into the future as a technological creative assist.

Will the Public Backlash Against AI Hurt Suno Long-Term?

You’ve likely seen the evidence that the public is starting to speak out against AI due to the fears of it becoming a massive super-intelligence. Other proof is out there showing a backlash against AI music in particular, even though it’s undeniable Suno has a record amount of users and listeners.

Last report showed they have 100 million users around the world, not including an evaluation at $5.4 billion. A good chunk of the world is either using the app or going in to listen to music produced there. Perhaps the creators signing up are creating while secretly abhorring what AI is doing in music. Others (like me) possibly realize they could save money using Suno as a music production studio while still writing their own songs.

No lists are available showing exactly what parts of the 100 million users are doing there. One thing we have to admit, though, is that many people continue using things they feel guilty using. It doesn’t have to be said what those might be, though Suno may be a guilty pleasure for far too many people. In fact, some may just use it for personal entertainment rather than thinking they can make money off their songs.

The numbers still show there’s some level of acceptability of AI in the world of music. Whether it continues in a time when AI CEOs are saying AI could eventually run amok and kill off humanity before the end of the decade is worth looking at closely.

As long as AI can stay as business and creative assists only, it’ll have already helped hundreds of millions of people accomplish things without taking away personal creation or ambition. Suno, with its new licensing agreement, also shows personal responsibility in protecting IP, despite there always being an argument on whether reusing past creative works really amounts to fair use.

I’ll continue to use Suno to create studio recordings of songs I write myself. Recently, I was preparing to find alternatives in the event Suno would shut down before the year was out due to its legal problems. Now it’s clear they’ll be around, and may end up being the most profitable survivor as Hollywood banishes AI, and AI business assistants become a little too scary to use.

All that, unless Suno’s AI bots start going rogue and end up screwing over those who use it merely to avoid studio production costs. I’ve noticed the more prompts you fed it on Version 5.5, the system would basically rebel and ignore your changes to go its own way. While some might say it could improve a song this way, going off the path of our own creative directions could signal a future Frankenstein’s monster scenario at Suno as well.

In Part 12, I’ll look at all the brouhaha over AI agents possibly destroying humanity someday. What are the chances of AI shutting down essential functions of your job within a short time—hence disrupting your job and career already?

/End

Part 10: My Experience Using Suno Pro to Create a Studio Quality Single of a Song I Wrote

Featured1 Comment

One of my creative pursuits outside of writing (and a 14-year freelance writing career), was music performance and writing songs. I was even given a chance to be a piano accompanist to a singer on the road once, but I turned it down to avoid living in hotels (no, possibly motels) every night. Regardless, the passion for writing songs that could be creative, yet still mainstream, went on. Much of that stagnated during my busy writing career years. Only in the last couple of years have I seriously gone back to finally realizing some songs swimming around in my head.

The irony here is that I’d just started this new blog theme about AI and human creativity while finishing up a new song from last year. I was hoping to find a way to create a studio-quality version of the song, yet didn’t have the resources to hire out musicians and record a demo in a local studio. Despite having a work colleague who leases a local music studio, it still would cost considerable money to hire a singer and musicians for possibly several days of recording sessions.

Enter my curiosity with Suno, particularly their Suno Pro tier. I thought to myself, why not just dive in, sign up for Suno Pro for $10 per month, and then create a studio version of my song? It was a test to see whether my initial aversion toward Suno was perhaps too much of an impulsive negative response.

What happened afterward changed my view on the usefulness of Suno Pro, even if I still have reservations about how it turns your songs into ear-popping studio magic.

You’ll Be Shocked at the First Suno Pro Results

We all know that many AI bots are a bit, what shall we say—(stupid?)—in realizing our prompts. You really have to be overly specific or use key language to get many AI bots to accurately create what you want. The problem there is, you shouldn’t become a lethargic creator and let AI bots create everything for you. It requires a lot more in explanation if you want it to create everything from scratch. In my case, I just wanted a studio-quality take of my first song called “Whispered Thoughts (The Girl Who Reads My Mind).”

All I did was write one small paragraph of prompts to make Suno Pro create a studio take with a male singer. I didn’t need anything else since I write all my songs organically at the piano on the Notion composing app. Yes, I use an Apple Pencil on my iPad Pro to write it all out one note/chord at a time—with my own lyrics.

The most amazing thing about Suno is that it seems to get what you want right from the first take. When I played back the first cut with a male singer, I was impressed at hearing the song fully realized with studio sound. This was after months of just hearing a piano solo playback on the Notion app.

But then, I decided to create some alternate versions. Since my above song is intended to have gender fluidity in the lyrics, I created a female take next. It almost sounded like Taylor Swift singing, yet it turned out more than well. However, it wasn’t quite as perfect as the initial male take. After several more alternate takes with both male and female singers (including an attempted duet), Suno seemed to kind of go off the rails on making it really radio-friendly, even if the prompts hardly changed.

Perhaps Suno should be called “One Take Suno” since it appears to give you the best possible cut on first try. All other versions seem to variate too much off the best path to a point of annoyance. Once that happens, it’s hard to get Suno to go back to the original style.

My Intentions for the Song

Thanks to the mostly stunning results from Suno Pro, my intention is to take “Whispered Thoughts” and place it on SoundCloud through their basic Artist tier. There, they’ll promote the song to the right audiences, and then it’ll stream on their site for a small price. But my full, long-term plan is to pay into a streaming distributor in the next several months and get the song out to many other streaming outlets. A distributor like DistroKid can help on that front, not including joining a publishing watchdog group to help with royalties on other media uses. Also, joining performance rights organizations like ASCAP or BMI brings additional eyes to your songs and pays you when it’s played somewhere.

All of this goes on the idea that the initial singers will not be me, but digital creations. At the core of importance, though, is continuously noting that I’m the complete writer of the songs. I’m still an adamant proponent of creating things from your own brain rather than letting AI constantly tempt you into creating something for you. Never will I let Suno create a song for me—even if some will argue that letting it create a studio cut is at least a quarter of an AI creation.

I argue not, as long as you specifically tell Suno not to deviate from the music. You have to move the dials a bit on the “Options” section when putting in your prompts. They let you choose how much of a percentage of Weirdness and Style Influence you want based on the recording you feed it. In my case, I uploaded a video of the Notion app piano solo as the initial source.

I’ve Come Around to Accept AI as a Creative Assistant

My first impressions of Suno from the outset were one of derision before putting it to use. You have to be suspect at first when you realize all of the singing voices and instrumentation used to create a studio cut are taken from existing sources. This bothered me since I’m a staunch advocate for protecting IP and not having it being made available for free license.

Those who argue that a large potpourri of these musical properties constitutes copyright infringement may have to change their philosophy. I’m not saying I have, yet it’s a saving technological grace to be able to full realize a song in studio quality to save time and money. The fairest thing that could happen is for Suno to pay a small fee to the IP they use to help musicians and songwriters. Considering Suno is likely already making a bundle just on subscriptions and community use, they could arguably afford that.

Courts are still trying to mull this over as lawsuits pile up. One could argue that when existing IP is stirred into a large digital pot, it becomes something else. When it helps a songwriter like me turn a written song into a professional studio version, it’s actually becoming a reciprocal process in helping musicians.

This is why I now consider Suno as a worthy creative assistant and nothing more. Also, I fully intend for these Suno-made studio cuts of my songs to be promotional tools in enticing a real (hopefully notable) singer to record them. Not that you can’t just sample your own voice into the system and turn the singer into something sounding like yourself. I just haven’t done this yet since the first take with the male AI singer was right on the money.

I’ll promote the song on this blog once I get it onto SoundCloud. After that, I’ll be looking into DistroKid, ASCAP or BMI, plus a mechanical rights organization (or an MLC). This entire business side setup is the new way forward if deciding to stream your own songs on streaming outlets and protect your work.

My Future Use of Suno Pro

Other songs are in the pipeline from yours truly. This includes a Christmas song that’s about done at time of this writing. That one will be a mix of Gospel-Rock and Hip-Hop. It’ll officially be my second single to group up with “Whispered Thoughts” on streaming outlets this fall.

Two other songs should follow before end of the year. One, a Country-Blues song, and the other with a Dance-Pop vibe. Latter will be a new lyrical approach in having an AI bot lamenting a relationship in a love song.

Suno Pro will probably be used for all these in creating studio-worthy versions. My only reservations are the mounting lawsuits against Suno for using existing music and vocal properties. Should it start mounting even further, it’s possible Suno would have to shut down—or I’d look into other options.

From now on, though, I plan to envision AI as a helpful assistant rather than assume everyone will use it as a complete creative crutch. Not that I’ll completely back down from the potential dangers of AI in taking away from human endeavors. The problem is, when handed everything, humans are going to become lazy and let the bots take over as much as possible. Many corporations are doing this now, and it doesn’t always go over or end well.

In the realm of songwriting, it’s all the more important for a human to create songs and bring out the depth of that artist’s soul. Just try letting Suno create a song for you, and you’ll immediately be able to tell the difference. That’s especially true in the realm of lyrics, something I write and hone as well.

I go forward now with the theme of this blog still intact, but with a better formed philosophy built on where AI should go and where it needs to be cut off. Whether Suno continues being a part of that remains an unknown as the complications of IP law and other opinions from other users still pend.

In Part 11, I’ll continue my look at Suno and a recent lawsuit against them from German company GEMA in the app using copyrighted recordings for AI training. It could lead to fast changes in how even AI-assisted songs are viewed or put together.

/End

Part 6: Is AI-Made Music Irresistible to Humans, or Being Consciously Shunned?

Featured1 Comment

No doubt some of you (and many others) never thought AI-produced songs would become big hits on influential music charts. And yet, there’s been two AI-made songs that went to #1 on two fairly big charts in the last six months of this writing. One is “Walk My Walk” by Breaking Rust on Billboard’s Country Digital Song sales. This one alone is considered the first instance of an AI song hitting numero uno on anything owned by Billboard. The second one is “Celebrate Me” by IngaRose on global iTunes charts at the beginning of this year.

While no AI song has made it to #1 on any mainstream Billboard chart yet, it doesn’t count others that came close. Various others have hit the Top Ten over the last year, albeit on niche music charts and, again, not on Billboard.

Still, the news of the ones charting that high made big news on social media. It made even skeptics begin to speculate on what could happen if AI creations continue making headway in the mainstream music industry. But, it may also send a new signal at what’s soon to happen based on a recent report saying music listeners are beginning to reject AI music creations.

This hasn’t stopped people from making AI songs. There was even an instance of a songwriter who reported that his unfinished song was stolen by an AI creator, completed, and posted on a streaming service.

It’s news like this that might be the final straw for music creators and fans. But it’s time to look at the reality of whether the music industry can control AI taking over things (like the Oscars now will for movies), or if it’s too late to prevent a turn of the tide. The outcome may involve the sheer will of the public at large.

Is There Really a Soul in AI-Produced Music?

I’ll admit that I listened to the songs listed above for research, and they were not bad. The AI artist IngaRose, in particular, had a voice all music industry labels and producers look for in ways of standing out. But, reality is reality. A digitally produced voice still has a way about it that just feels empty when listened to more intently. It’s everything to do with sincerity in whether you truly believe in what the artist is singing about. There’s a humanness to sincerity in singing that even far too many human singers don’t have.

Beyond that, there’s also a new psychological aspect to AI singers—in that if a listener is told it’s AI, they’re more apt to feel…well, cheated. It seems your fellow human ultimately wants fellow humans to express themselves in music as a form of an artistic gift.

The above comes from a recent study citing a decline in listening to music produced by AI, most notably in the coveted Gen Z and Alpha categories. As reported, it seems there’s a chain reaction to that in some major artists penning an open letter saying “Say No to Suno.” With the latter AI software being a leader in producing AI music, it’s possible avid fans of those artists agree with the AI issue and began turning their backs.

Others are saying AI is even using cultural appropriation, as in black music being digitally generated while mimicking fake life pains. Yeah, it’s a reminder the real human struggle is where the true magic is in the best music produced over the last 100 years.

Worst of this is the knowledge that AI apps like Suno are also taking away from royalty rights of real artists. With many listening to music to inspire creating music of their own, the realization that AI could take that away (along with basic jobs) is a beacon for a possible approaching backlash.

So, I pose a bigger question: Are we going to see AI back off in the music industry due to much of the public hating it? What that looks like is also in question as AI corporations continue to push it hard on the public, whether people want it or not.

What Would Happen with an AI Music Industry Pushback?

The biggest problem isn’t so much how much the public rejects AI in music. It’s in how far AI companies continue to push it out there as a piece of digital candy. Even if some people reject AI-produced music, it’s going to keep happening as the technology gets better and there’s pirates out there willing to use it for maximum moneymaking.

We could see an eventual public division in the world over rejecting AI and those who embrace it all the way. And that almost looks biblical in scope if you consider how much AI could take over everything, including running the world in various significant ways. And it may, again, come down to the very point of this blog: Can human creativity prevail rather than be fragmented to its own minority existence in our world?

It all comes down to the temperament of people and what they do. In the immediate term, there may be an AI burnout phase where people reject creative things done by AI. Once AI becomes more refined, it may be too good to resist for a large amount of our world population.

Regardless, most people seem to be reawakened to the very idea behind music of all types. And that is, human pain and expression are what make good songs/performances. There’s an aura there people can feel that’s almost ineffable, yet clearly present. The palpable science behind this may be the next big discovery to prove what makes a human creation more compelling.

Many people also don’t want to be cheated of that aura. Perhaps they’re still being fooled by it—then changing their minds on support when told it’s AI. It’s why it’s important now that streaming services officially tell us what’s created by AI and what isn’t. Not all are, so AI continuing to encroach on the music industry would have to be a great deception to continue growing.

Don’t be surprised if said deception grows larger when big money is at stake. The good news, at least, is that (through the above study anyway), humans want real humans to express themselves musically. If their bullshit detectors are always on full force, we’ll continue to hear real performers creating their own songs and making awesome performances.

Yes, but is the Music Talent Pool Growing Thin?

Sure, the argument can be made we don’t have the type of musical artists we once saw. Those who get pushed to the mainstream are sometimes mediocre and don’t express themselves as compellingly as prior generational artists did.

In this case, AI can create faux artists who bring that type of expression back. I’ll admit the AI artists I heard seemed to express themselves deeper, even if clearly infringing by being an amalgam of great artists from long ago or the recent past.

What’s needed is supporting the artists out there who really express themselves with sincerity in a song. Far too many don’t get the support they need by the public due to lack of promotion. With the threat of AI taking away those individuals, the best thing that can happen is just more acute awareness of would could happen. Plus, since millions of real people would love to make legit money in the music business, the sheer will of the public may be the future—uh, AI unplugging.

In Part 7, I’m going to look at how AI creative assistants seem to be imitating a mix of existing content available online. Is it really copyright infringement when it’s a mix of everything ever created?

/End

Part 2: What Will Become of New AI Actors and Musicians Encroaching the Arts?

FeaturedLeave a comment

Most of us would probably admit that we once scoffed at the idea of AI musicians and actors ever becoming a reality in the arts world. Well, what seemed impossible when we became so incredulous merely two years ago has become real. Now we see at least one AI actress (named Tilly Norwood) making headway as a real digital actor. While we haven’t yet seen evidence of how effective she is in long-form content, Norwood’s creator made a multi-million-dollar deal.

It’s becoming much more undeniable in the music world with a number of AI singers/bands currently having hit songs on the charts. You may have heard about Xania Monet who’s the first AI-made female singer to have a Billboard chart hit in the genre of R&B/Gospel. There’s even an AI country band called Breaking Rust that had a #1 song recently. The Christian market has also succeeded in creating an AI band called ChildPets Galore.

A few other AI artists are making headway of late as well, giving concern in the music industry to what constitutes realness from fakery. It leads me to this very blog and trying to predict where this artistry trend goes. Do people really care now if an artist is AI, or are we just not looking/listening well enough anymore?

The Half-AI/Human Creative Deal

If there’s any benefit to having AI actors and musicians, it’s that they’re created by a human being. The content being created for the AI creations is being done by humans as well, so it’s merely in the presentation. With that, you could say it’s the purest example of AI working alongside humans. Yeah, the human can benefit financially while not having to be in the limelight. Or, that is, until the AI creation becomes fully sentient and wants a cut of some (or all) of the profits.

Much of this also sounds like a Frankenstein’s monster analogy come to life. And, it distorts how we enjoy art since the point of acting was to tap into the human condition so the audience relates. The same goes with the art of singing a song since the human soul understands life’s pain better than any AI creation ever will.

Hearing pain and real life lived in a song is what made the names of legendary human singers of the past. Great singers such as Billie Holiday, Frank Sinatra, Whitney Houston, and Mariah Carey have all proven that in the last 90+ years. Of course, many of those great singers are now passed, and even Mariah can’t sing the high notes as she did 30 years ago. No doubt about it the public wants to hear great voices again.

Having heard Xania Monet recently on Apple Music, I’ll admit she has amazing digital pipes. However, I noticed she sounds the same on all her songs, showing there isn’t a multitude of dimension to her voice. It’s simply because there’s no real soul there or aura a real human uniquely exudes. I’m willing to bet that AI will never be able to convincingly duplicate a human voice to the point where it fully understands or exudes the complexities of life.

Most surprising within the AI musicians list are the Christian and Country genres. Considering these categories demand the most human aspects to conveying life’s hardships and the Holy Spirit, having reported AI musicians singing songs there is an eye-opener. It raises bigger questions about whether the emotions these faux singers conjure is really the ultimate in deception. But what I want to know is how people really feel once they know these singers are AI-created.

Deterioration in true listening ability may be to blame in the Country and Christian markets becoming encroached by AI singers. It may also come down to perception. Reportedly, a lot of AI singers are popping up on the charts that the listening public don’t realize are AI. Many streaming platforms are supposed to be filtering out AI artists, yet it seems they’re letting them slip through the cracks due to the apparent proof they can make fast hits/money.

The Hollywood AI Debate

It’s a lot more complicated in Hollywood regarding the emerging trend of AI actors. Coalitions are starting to be built there to ward off too many AI creations taking away jobs from human acting talent. And that may help curb the tide for a while. That is, unless the AI actors being experimented with end up advancing and start winning “awards.”

At issue here is Tilly Norwood, the only fully-realized AI actress who may be starring in upcoming movies. The company (Particle 6) and person behind her creation reportedly made a deal with a talent agency to eventually use Norwood in some kind of cinematic context.

Thanks to SAG-AFTRA, she isn’t recognized as a true actor yet. And Norwood hasn’t been used in any movies other than a short film posted online. Most of her is just seen on her own social media account and other online appearances. While she does look impressive, her Dutch creator (Eline Van der Velden) is going all out to get her creation on the big screen. Based on the short film Norwood starred in, you can still tell (if you look closely enough) that there isn’t an aura present.

Yes, this makes me fear Norwood soon becoming self-aware and being capable of invading my tech space eventually. The worst sci-fi scenario is lambasting an AI actor and having them invade your smart refrigerator to retaliate.

AI musicians are clearly going to be taking over faster than AI actors because the former can fool people with sounds vs their physicality. All it’s going to take, though, is Norwood being in a hit movie to move the needle. Some AI creators are way too much like Dr. Frankenstein in willing something like this to happen. BTW: See Guillermo del Toro’s recent “Frankenstein” as the ultimate media metaphor for all of this.

The Ability to Scope Out Human Art

Perhaps part of the problem here is that older music and movies aren’t being watched or listened to as much as they should lately. A deep exploration into what made music and movies great in past decades proves that humans brought a broad depth of emotion and understanding to the human condition. Newer talent arguably haven’t done that quite as well, with those who do being relegated to the realms of small, independent productions.

Or, maybe it’s because we want actors and musicians to bring the utmost in emotion, something many artists just can’t tap into anymore. Even der Velden above argued her Tilly Norwood creation would help make it “safer” for actors who sometimes go to the psychological depths to prepare for complex roles.

She misses the point, though, that it’s all about the compelling essence of a real actor. They don’t have to do dangerous things to mesmerize us with a performance—assuming they have the skills.

If humans are losing those skills, then we need to find them again by listening and watching what came before. AI may eventually learn what those skills were, perhaps making them look brilliant to those who have no idea what they missed with the humans.

As usual, the past and the future are probably going to intersect more than ever to bring any ethical sense to what AI should do.

In Part 3, I’m going to address the recent panic over AI agents talking to one another on their own social media site called Moltbook. Are humans pranking us with AI sentience, or is it the beginning of AI autonomy on everything?

/End