• The cover of the 'Perl Hacks' book
  • The cover of the 'Beginning Perl' book
  • An image of Curtis Poe, holding some electronic equipment in front of his face.

AI-Generated Content: Innovation or Intellectual Theft?

minute read



Find me on ... Tags


Introduction

It’s no exaggeration to say that we’re worried we might lose our jobs to AI. The value of human language translation outside of specialized contexts is effectively zero. Call center workers have, for years, been replaced with AI systems, systems which are getting even better. Copywriters and journalists are seeing generative AI encroach on their work.

And the ability of generative AI relentlessly advances, though not with the same growth in ability we saw in the early days. Many of the advances are subtle, but powerful. The latest version of ChatGPT, released as a preview a few days ago, shows some amazing abilities to plan.

Now Billie Eilish, Nicki Minaj, Stevie Wonder, and others call for protecting their jobs against AI .

Jobs versus Progress

Do you remember the public outcry when the availability of cheap alarm clocks put knocker-uppers out of work? No? You don’t even know what knocker-uppers are?

OK, what about when electric lights put lamp-lighters out of work? What about agricultural workers, switchboard operators, ice cutters, toll booth collectors, or, or, or ...

There was no outcry. Those were jobs for poor people. Often immigrants. Nobody cared. “Progress!” was the rallying cry. But when famous artists call for protecting their jobs against AI, now there’s public outcry. Can’t have the rich and famous getting hurt, can we?

Of course, there have been a few outcries before, such as when Luddites were losing their jobs. I’ve written briefly about them before, but this is the exception, not the rule.

So there’s more than a whiff of hypocrisy here. It’s OK for poor, unknown people to suffer, but not rich, famous people.

Or, to give those complaining the benefit of the doubt, maybe they didn’t care that “Progress!” ruins the jobs of poor people because eventually they’ll find new jobs, but when technology can take everyone’s jobs, that’s where the fear comes in?

But that’s a bit of an extreme position, too, because AI is clearly not going to take everyone’s jobs; I’m not going to a robot barber any time soon. Currently we’re seeing labor market transformation, not collapse. Some jobs are being lost while others are being created.

But really, it comes down to one thing: the scale of the change we may be facing. Before we get to that, let’s talk about the electronic elephant in the room, AGI.

Artificial General Intelligence (AGI)

Definitions of AGI vary because we can’t define “intelligence,” but the definitions revolve around when AI might be able to solve all or most intellectual challenges at or above the level that humans can. But for the sake of argument, we’ll just say AGI means “equal to an average human.”

So let’s consider the rumor that OpenAI is considering charging $2,000 a month for their next generation model . We have no idea if those rumors are true, but it sounds insane, doesn’t it? I pay twenty bucks a month for ChatGPT. Why would I want to pay one hundred times that amount?

If we hit AGI at some point, $2K/month is a steal and businesses will be scrambling to pay it. Why? That’s $24K a year. In 2022, the average annual US salary was about $64K . So if we have AGI and it can replace you, one AGI is about a third the cost of the average American.

Oh, but it’s even cheaper than that! AI doesn’t sleep. It can work a 24 hours a day, seven days a week. So the cost of the AI is an order of magnitude cheaper than your average American. And they don’t complain. They don’t stop. They don’t go to HR. They don’t join unions. They don’t require health insurance.

If AGI is achieved, what’s a business going to do? Their direct and indirect labor costs are an order of magnitude cheaper, and many of their continent costs go away. No more wrongful termination suits. No workman’s comp claims. No severance pay. No compliance costs. Unless there is significant legislation coming down the pike to stop this, any business which can adopt AI but chooses not to is going to be remembered fondly as naïve idealists who went bankrupt.

The Current State of AI

But are we there yet? OpenAI’s next model, Strawberry, allegedly is supposed to be incredibly powerful .

Strawberry will be able to solve problems and tasks that are beyond the capabilities of current AI models. It will be able to solve math problems it has never encountered; perform high-level tasks like developing market strategies and solving complex word puzzles; and perform “deep research.”

We’re not done there! Strawberry is rumored to be used in training Orion, OpenAI’s model after Strawberry, and the successor to ChatGPT 4o .

For those not working regularly with AI, it’s hard to understand the impact of our current foundation models, but they’re already transforming many businesses. If Orion-class models have a similar upgrade in capability—something we can’t possibly know until they’re released—all bets are off. If I had to guess, I think Orion won’t be AGI, but we might be bumping up against it. I am not willing to wager large sums of money on my guess.

The Legal Landscape

OK, that’s some background. What about the creators who don’t want to go the way of the knocker-upper and lose their jobs to AI? There are numerous lawsuits against companies training models on artist’s works, but it’s unclear what will happen.

What is clear is copyright law: I have every right to copy your style, so long as I don’t copy your content, or pass my work off as your own . I can paint like you, I can sing like you, I can write like you. (Well, I can’t, but you get the point).

It’s a critical concept because ideas can’t be copyrighted. Imagine all of the lawsuits suing writers of books because similar plotlines have been used (boy meets girl, girl hates boy, she changes her mind), and then you, like the original author, used an unusual second-person POV. It would be madness.

How could I copy your style? By reading your writing. By viewing your artwork. By hearing your songs. That’s what much of generative AI is doing. Somehow, what you and I are allowed to do—see things and imitate them—generative AI is not allowed to do.

So it’s not clear that artists can claim copyright protection for AI-generated output. In fact, since anyone can produce great art, whether images or songs, via AI, the Copyright Office would be overwhelmed by trying to process all of those requests. We’re simply not set up to handle this.

Initially, the US Copyright office denied copyrights for all AI-generated work. That seems odds because I’m allowed to copyright art created with a guitar, a brush, or Photoshop, but not Midjourney? Well, the Copyright Office started relaxing its opposition last year , so long as there’s substantial human creativity involved in the process.

But yeah. What does that mean? On its face, the artists probably aren’t going to win on the grounds of copyright violations, though there are many pending court cases surrounding this .

The arguments seem to largely fall into two categories. The most prevalent is that various companies are “illegally” scraping copyrighted content to train their models. Well, how do I learn something? I look at it. I study it. I’m allowed, but the AI is not? That seems a curious dichotomy.

On the other hand, if the content explicitly forbids such scraping, then yes, I can see the plaintiffs prevailing—if they can prove that their content was scraped.

But there’s another curious angle. Many cases are arguing that defendants have made copies of copyrighted work for training their models . In fact, you should probably read through that entire link to get an idea as to the scope of these lawsuits. If it can be successfully argued that defendants are illegally storing lossy, compressed versions of copyrighted works in the weights of models, maybe there’s a case? Though it’s worth pointing out that you keep lossy, compressed versions of copyrighted work in your brain. Carbon good, silicon bad? That being said, humans do so at a much smaller scale than AI.

I know, it sounds like I’m defending AI at all costs, but I’m not. Bear with me; I’m trying to be fair. People I talk to about this tend to come down on one side or another, reflecting their personal views, regardless of the merits of a given case. Of course, we tend to do this about many court cases, facts be damned.

Despite not being a lawyer, or having anything resembling a law degree, I’m willing to bet that the AI companies will largely prevail on these issues, despite losing on some points. I say that because I went to university to be an economist (came out a software developer) and I know what often happens when large sums of money meet murky legal terrain: money wins.

Legislating AI

So the law needs to change, but how?

Let me put it this way: any legal change you suggest is going to have wide-ranging consequences, not all of them good. For example, famed Perl hacker Randal Schwartz, was convicted of a felony for doing what he thought was his job .

At the time of the initial case, Oregon law, title 16, chapter 164 read :

Any person who knowingly and without authorization uses, accesses or attempts to access any computer, computer system, computer network, or any computer software, program, documentation or data contained in such computer, computer system or computer network, commits computer crime [is guilty of a class C felony].

You could, in theory, be guilty of a felony for leaving a voicemail for someone who told you not to call them. Or maybe a family member shared their Netflix password with you. That’s a felony! Or you use a VPN that allows you to read US websites from outside the US. That’s a felony!

That law has since been amended . It was well-intentioned, but the larger the scope of a law, the more possibilities it can be abused.

Remember Prohibition? Though well-intentioned, it almost single-handedly created organized crime in the US .

Remember the Smoot-Hawley Tariff Act of 1930? That was passed during the Great Depression, raising tariffs to protect US agriculture. Due to the retaliatory tariffs, this law was blamed for worsening the Great Depression worldwide .

Remember the various three-strikes laws? One man was sentenced to life in prison for shoplifting a pair of white socks worth $2.50 . Others received life sentences for stealing food. But the law was well-intentioned.

Conclusion

So where does this leave us? We’re standing at the precipice of a technological revolution that could make the Industrial Revolution look like a warm-up act. AI may be coming for our jobs and it’s not just the assembly line workers or the telephone operators this time—it’s gunning for the artists, the writers, the coders, and maybe even the CEOs .

Is it hypocritical for famous artists to cry foul now when we’ve collectively shrugged at decades of technological job displacement? Perhaps. But maybe their star power is exactly what we need to finally have a serious conversation about the broader implications of AI on our workforce.

The legal battleground is murkier than a swamp in fog, with copyright laws struggling to keep pace with silicon’s relentless march. Some are calling for AI companies to pay licensing fees for the copyrighted material they use to train their models, similar to how streaming services pay royalties to artists. Sound fair? Plot twist: Such a move could inadvertently raise the barriers to entry in the AI field. We might end up with a greater market consolidation, with only the tech giants able to afford the licensing fees. Is trading artist exploitation for tech monopolies really a step forward? Or are we just rearranging deck chairs on the Titanic of our economy?

One thing’s for certain: the AI train has left the station, and it’s picking up speed. We can either scramble to lay down tracks in front of it, or we can start thinking about redesigning our entire transportation system.

The real question isn’t whether AI will take our jobs—it’s what we’re going to do if it does. Will we finally address the fundamental issues of wealth distribution and the value of human labor? Or will we continue to bury our heads in the silicon sand until even that job is automated?

The future’s coming, folks, whether we like it or not. The choice is ours: adapt, evolve, or become as obsolete as a knocker-upper in a world of alarm clocks. Welcome to the brave new world—hope you brought your resume.

Please leave a comment below!

Full-size image


If you'd like top-notch consulting or training, email me and let's discuss how I can help you. Read my hire me page to learn more about my background.


Copyright © 2018-2024 by Curtis “Ovid” Poe.