Where We Stand With AI-Generated Art in the Current Moment: 2025 Edition

It’s that time again. Let’s dive back into a deep conversation about the complicated world of AI-generated film and photography.

In September of 2022, I wrote an article for Fstoppers entitled, The Coming Menace of Artificial Intelligence and How We Can Respond As Artists. At that point, the general public was just becoming aware of names like Midjourney, DALL-E, and ChatGPT, and the technology was seen as little more than a curiosity. But, as a working artist who makes my living from my passion, it was clear to me that this little curiosity was merely the tip of the iceberg.

I had immediate reservations about AI—two, specifically—both of which remain unresolved. One was the obvious question of artists versus the machine. Could a tool promising to generate realistic-looking images in an instant pose a business threat to photographers who make a living creating images in real life? The second, a still more galling question to me, is the fact that these AI systems’ abilities to generate such realistic results are fundamentally based on the widespread theft of intellectual property from human artists without their consent. It’s one thing to have a new competitor on the field. It’s wholly another to have your life’s work stolen without recompense, then be asked to compete against yourself, essentially, in an already crowded marketplace.

The concept of adequately compensating the artists whose work was stolen to train the AI models remains, for me, the primary moral question within our march toward the new era. The question of whether or not AI will ever be able to appropriate the feel of real photography or suitably simulate the emotion of a real human being is really more a matter of when rather than if. Essentially, the tools that currently exist have advanced to the point now where even a trained eye might have trouble telling it apart from human-generated content. If we really want to nitpick, it’s possible to argue that AI-generated material is still not there yet. But what is not even remotely in dispute is that, in the nearly three years since I first wrote about AI, the levels of what you can accomplish with it have grown by leaps and bounds. I’m not arguing that the system has reached perfection. But comparing it to what was possible three years ago, I think it’s safe to say that the technology is advancing faster than any other technological innovation in recent memory and will reshape society in much the same way as the Industrial Revolution, the birth of the internet, and the advent of social media.

My journey with AI has been both begrudging and gradual. Even when I wrote about the coming menace of AI three years ago, I was concurrently training myself on how to use the technology. Even if one views it as the enemy, it pays to know your enemy. In my initial tests three years ago, my goal was to try and use the technology to create still images similar to the ones I would shoot with a traditional camera. Then, I moved on from that to try and make those images move in a way that was realistic and could potentially rival something I would shoot with a motion picture camera. The next frontier was sound—specifically, trying to find a way to sync dialogue to picture, which really took some doing.

In fact, my latest tests with AI video have been largely driven by the release of two things: Google’s Veo-3, which allows you to create AI video with sound and voice already attached, and Midjourney’s new video generator. There are a number of other systems on the market that I have combined over the years to try and make AI video. But I think the release of Veo-3, especially, will go down as a watershed moment as we look back on the development of this wave years from now. Not because the system is perfect—I’m only using the free trial period now and my results have been… bumpy. But the reason I say its release will be a watershed moment is because it fills a key market opportunity in what is currently the wild, wild West. Right now, in order to really create an AI asset you are happy with, it often requires you to access multiple different AI platforms to perform a multitude of tasks. In other words, it's a labor-intensive process within an AI world whose main promise is efficiency. Veo-3 being able to automatically link sound and video will streamline the process for many casual users and bring even more customers into the fold.

Of course, as a human artist that needs to compete with the machine, this is awful news. But from an economics perspective, this product benefit makes the platform more appealing to an expanding market and will no doubt spawn more copycats among Google’s competitors. Like I said, Veo-3 still seems to have some limitations. But, again, as they say, today is the worst the technology is ever going to be.

Case in point: I recently decided to do a remake. If you’ve known me for longer than five minutes, you’ll have heard me rail about how much I hate remakes and sequels. But, in this case, this remake was made for a more exploratory purpose. In 2023, when I was first experimenting with trying to get my AI images to move, I began making little short teaser videos of what was possible. With technology having grown exponentially in the last year alone, I decided, as an experiment, to remake one of those old films using modern AI tools and techniques.

The results were staggering. I don’t mean to say that the result is the greatest AI film ever made. And there is still a hefty dose of AI weirdness that pervades the process, which even my editing skills aren’t completely able to disguise. But if one were to look at the results back to back—the original 2023 version followed by the 2025 reboot—it is absolutely undeniable how much has changed in two years. Project those improvements ahead another year from now, and you start to get the full picture of where things are going.

So, what does that mean for us as artists? Should we all trade in our cameras for laptops? Is humanity a thing of the past? Are the industries we’ve dedicated our whole lives to doomed to a slow demise? Honestly, I don’t think so.

I know you want me to say that AI is just a fad and that people will rise up and reject it, but I’d be lying to you if I echoed those words. AI’s effects on artists, which is the topic of this article, are merely a minute portion of the larger picture. AI art gets a lot of headlines. Artists are always held out front for either praise or as cannon fodder. But the impact of AI is being felt way beyond the realm of the art world. It's transforming science, business, and healthcare. It’s even changing warfare (not for the best). I was actually listening to the godfather of AI, Geoffrey Hinton, speak recently, and he made an excellent point about the current and potential uses of the technology by the world’s military and how much money is already changing hands. If we know anything about the world, we know that the rich want to get richer, and AI is a potential goldmine for industrial complexes around the world in multiple industries with far deeper pockets than independent artists. So the tech is going to continue to advance because there are people with vested financial interests in seeing it grow.

At the same time, there are some very legitimate benefits: the potential to better sift data that might help cure diseases, the ability to spot trendlines that might help improve our society or make things safer. Keep in mind that both of those benefits rely on human beings using the technology ethically, and that is still a question left to be answered, but there are real, tangible benefits that could come from the technology. Thus, again, AI is going absolutely nowhere.

So how do we learn to live with it? Well, my initial fear when AI was released was that it was going to completely replace human artists and wipe out 90% of our careers overnight. After spending the last three years testing it, learning what’s under the hood, and gauging the public’s reaction to it, I’m currently of the mindset that AI art will exist alongside human art rather than replace it completely. There will, no doubt, be a major economic impact on the marketplace. While I don’t think human artists will be eliminated, the influx of another art form will still further squeeze an already oversaturated market. The overall size of the market itself won’t grow. Rather, the arrival of AI will just siphon off a large section, leaving human artists competing over an increasingly smaller piece of the pie.

Think of it like this: At one point, advertising was all hand-painted billboards. Then, photography came in and the number of hand-painted signs went down while signs featuring photography took over part of that load. Eventually, you got moving video billboards. Those took a chunk out of the market away from photography and painting. Both still existed; their slices of the pie just got smaller. Now, you are going to introduce AI into the mix. Painting, photography, and film won’t be shuffled off the board, but they will likely be downsized to a smaller piece of the pie to make room for AI-generated material to have its piece. I do think that, as long as there are human consumers of art, there will be a market for human creators of art. But the economics of supply and demand will no doubt affect the type of art that is created, the budget levels at which it is created, and how many artists are able to make a living in their chosen profession.

Now, whether that outcome is a positive or a negative is an entirely different essay altogether. But the fact remains that coexistence is the bare minimum of what is likely to happen. Three years after that initial article I wrote, we have transformed from the “AI is coming” world into the “AI is here and a permanent part of life” world. A multitude of questions still remain—both technological and moral. If AI has come this far already in being able to replicate reality, how far might it go in another three years? What responsibility do both the AI companies and independent producers have in ensuring the economic stability of artists (and their support teams) who will be displaced by a flood of less expensive AI offerings in the marketplace? And how do we properly compensate the artists whose intellectual property is the literal basis for the large language models (LLMs) but who have so far not yet been afforded legal protection? I actually have an idea for that last part, but, again, that is a topic for an essay of its own.

We live in a world where technological advancements are literally happening too fast for me to keep up with. I continue to have reservations. Not just because I am a born Luddite who still daydreams of the days before cell phones and the internet, but because there are many ethical concerns that still need to be addressed, including, but not limited to, whether these advancements really serve humanity at all. But, if we’re really dealing with the facts on the ground versus what I may or may not wish to be the case, AI is here to stay. It’s up to us to create a society where we can take advantage of its benefits, yet also provide guardrails to help human artists thrive now and into the future.