At the intersection of creativity and AGI stand some of the most pertinent questions in this field. If a future human-level superintelligence, an AGI (Artificial General Intelligence), will create original art, who would own it? How do we know it’s “original’? At what point do we say we are dealing not with a collection of algorithms but a real, sentient being?
In unraveling this, let’s look for a definition of what level of performance such an intelligence would need to achieve to qualify for an AGI status. After all, beauty (and intelligence) are in the eye of the beholder. In this case, the yardstick has been human intelligence. But who is to say we will understand the kind of intelligence an AGI would possess? Already, we know that the workings of ChatGPT are mysterious to its own engineers.
In their 2025 paper “A Definition of AGI,” The Center for AI Safety, representing nearly 30 top international researchers, defined an AGI “as matching the cognitive versatility and proficiency of a well-educated adult.” The group came up with a framework that broke general intelligence down to ten core “domains” that include reasoning, memory, and perception. They also applied this rationale to finding a benchmark against which to measure an AGI, which, of course, at this point is still just a potentiality.
If an AI makes a movie, or paints art, who owns it?
That’s the trillion-dollar question. Let’s say an AGI emerges and we all come to an agreement it has human-like (or beyond) capabilities. What then? Does it get to be a person? If it makes a work of art, does it get to own it and profit by it? And if it collects money, does it choose how to spend it?
The sheer barrage of questions underlies the significance of the issues to be figured out.
Can an AGI model be granted personhood?
At the core of the questions of ownership of a creative endeavor is the idea of personhood. How likely is it for an AI to be legally recognized as a person?
First, it should be said that there’s already precedent for something similar — corporations of today have limited personhood. But, under existing copyright law, authorship and ownership of creative works is reserved specifically for humans, not their machines.
An AI could create something really popular — we have already seen the millions of shares AI videos get on social media. Still, that won’t matter at this point, and ownership would be attributed to the people who paid for and designed the AGI. Subsequently, the profits would also flow to the creators and not the AGI.
It would take a major shift in our understanding of what a person can be to change this reality, but we are, after all, talking about a technology that has only been imagined. The question of AGI personhood has also been more academic than real.
Another obstacle: legal personhood implies having rights but also duties, liabilities, and “agency.” Current AI systems are far from meeting such criteria.
Having to take AI persons seriously might happen very quickly, however, as the recent case of Tilly Norwood showed. An AI creation, Tilly Norwood sparked debate when it received interest from casting agents.
If Tilly gets a part, who gets the money, the acclaim, the responsibility?
Even George Clooney commented on the commotion, sharing that it’s hard to become a star whether you’re human or AI. The situation hasn’t yet come to a head, which would happen when Tilly gets a part, but it seems that it’s only a matter of time before she or another AI actor achieves such a milestone.
AI in film: a case study
As someone who’s a part of both the creative arts community and the world of science and technology, I’ve come to use the so-called “weak AI” extensively in my work. It’s the kind of AI that is very helpful but hasn’t yet risen to intelligence levels comparable to humans. I’ll share a case study of what using AI to create is currently like and where it may be headed.
In particular, my work on the documentary film “Fish Fry & Fellowship” for the past three and a half years has become increasingly reliant on artificial intelligence. The film revolves around a traditional topic for a documentary, telling the heart-warming story of friends, both black and white, who come to a weekly fish fry at an African-American church in Palmetto, Florida. They meet to reminisce about football and eat fish in a community with a long history of segregation.
By this point, I’ve used over a dozen AI tools on this very human story — utilizing artificial intelligence to generate B-roll from the history of the era using Generative AI tools like VEO 3, Kling, Sora, Runway and MidJourney. I’ve upscaled to fresh and shiny 8K some grainy, low-resolution, or blurry images from decades ago using Topaz Gigapixel AI and Video AI. I’ve also used AI tools from Adobe to edit sound and an AI music generator Suno to create some period-appropriate background music.
From this cornucopia of tools, a story emerged unscathed in its humanity, while largely augmented by storytelling technologies previously impossible. What’s also emerged is a plethora of ethical implications that I’ve struggled to grasp, so many of them very new and not something I’ve ever had to consider.
As a filmmaker, I find the rewards of the AI tools too irresistible.
In some instances, I started out with original still photos from the 1950s and 1960s and animated them using AI generators. So the original images were very much real, but the resulting videos show an animation that wasn’t originally there.
The people in the still photos suddenly begin to move, at times smiling, at other times performing other simple functions that resulted from my prompts. I purposefully made sure to offer very simple commands that I didn’t think would affect the reality of what I was portraying. After all, these people that I animated did once move on their own, in their own realities. But their movement in mine wasn’t exactly their choice.
I also made some shots that were based on documentary footage or existing historical stills, but were still original AI creations. I used these types of shots to enhance the narrative, to add visual illustrations or poignant moments to what would otherwise be a bunch of talking heads.
The film is definitely helped by the techniques I employed. But what about the verisimilitude I am supposed to convey as a documentary filmmaker? Honestly, I think it’s still there. All that I’ve done is in the service of trying to capture the era I am portraying and making the stories of the folks at the fish fries I filmed more engaging and “real.”
But how real are they? You would be to right to question that. And I wonder whether, as a filmmaker, I now have a responsibility to label every shot derived in this way with a “made by AI” moniker so the audience isn’t deceived.
What about the use of recreated footage that has been a staple of documentary filmmaking? If it was ethically acceptable to recreate stories of the past using actors, what is different if the actor is made by artificial intelligence?
If anything, there’s an opportunity now to make that AI actor even more like the real thing, to make them live within the epoch they are representing. I think of these actors as very good animations. Is it my duty as an ethical creator to make sure my audiences are aware these are animations or does any of this really matter if the spirit of the story is correctly conveyed?
The true test will be from the audiences themselves. If they feel that uneasy feeling, experience an “uncanny valley” watching AI actors, and come to believe that somehow what they are seeing isn’t quite real, then my job as a filmmaker was not done well. At no point do I want them to be taken out of the narrative by other considerations.
But they may also spot nothing unusual and consume the narrative in the way I intended, undeterred by any doubts about the truth of what they are seeing. I see that as the most desirable outcome for me as a creator. In all my experiences with offering AI products to an audience, I’ve noticed one clear axiom — if the AI is good, no one will have a problem with it. If there’s something off about it, if it takes you out of the moment, people will notice and have a problem.
This leads us to the obvious consideration that most of us have already experienced on the internet — it’s getting harder and harder to tell what is real footage and what isn’t. This type of dynamic is ripe for abuse. Is it now the job of the government to save us from ourselves, or will the market be the arbiter of what is real from now on? Considering how fast this technology is developing, waiting for the government to catch up seems like a Sisyphean goal.
The same axiom of “weak” AI, I believe, will be the case with AGI-created feature films and AGI art — if it’s good and engaging, no one will care. And if the AGI’s art is intended just for other AGIs, then there’s even less to worry about. As we consider this, the question becomes: when will AGI Spielberg take the helm? How far away are we from that kind of singularity?
When will AGI actually happen?
In their pointedly-titled article “Will Humanity Be Rendered Obsolete by AI?,” Mohamed El Louadi of the University of Tunis and Emna Ben Romdhane of The Higher School of Commerce of Tunis, argue that the answer to their question is multi-faceted. What’s more, it may be the wrong question altogether. We measure intelligence by ourselves, evaluating our understanding of smarts against that of machines. But what if an AI that will become somehow sentient will develop an intelligence entirely unfamiliar to us?
They also propose, through a poll of AI leaders, that the arrival of AGI may happen through sudden, exponential growth. You won’t know what hit you. At least that is the belief of Sam Altman of OpenAI, Dario Amodei of Anthropic, Elon Musk, and Ray Kurzweil. They all see AGI being possible by 2030. Others, like the Oxford philosopher Nick Bostrom see AGI emerging close to mid-century, around 2050. Either way, it’s not in the distant future.
Join a community dedicated to responsible AGI integration. Sign up for our newsletter and stay informed on AI design and ethics.
Paul covered artificial intelligence and paradigm shifts in science and technology for Interesting Engineering, Big Think, and other publications. He’s also an award-winning filmmaker and educator.