Studios will steal artists’ faces and voices

AI-generated actors, singers and visual artists stand to replace the real thing

Jim Parker
Content image

(This illustration was created by Maclean’s art director Anna Minzhulina using the generative AI image program Imagine. Minzhulina spent weeks feeding prompts into the program, inspired by the essay.)

Jim Parker is a professor of computer science and art at the University of Calgary.

(This illustration was created by Maclean’s art director Anna Minzhulina using the generative AI image program Imagine. Minzhulina spent weeks feeding prompts into the program, inspired by the essay.)

Some folks worry that artificial intelligence will eventually make art that is better than what people can create, rendering human artists obsolete. I don’t think so. While AI may produce something surprisingly good once in a while, it will never be a serious competitor to the Rolling Stones, or Beyoncé, or any other performers and creators. Fine art comes from human emotions and experiences, whereas AI just averages a bunch of art pieces together and approximates them.

But of course, not all human art is great art. Historically, we have hired people to produce more prosaic materials like book covers, music jingles or graphics. That’s where AI will come in. Now, if someone wants a logo of, say, a robot riding a horse, they are less likely to pay someone to make one than to insert a prompt into a photo-generating AI like Midjourney and have it produce 100 results. If someone needs quick content writing for their business, they can just plug a few instructions into ChatGPT and receive cogent paragraphs. Even pornography is changing: there are many sites now offering nude images of people who do not actually exist, allowing users to specify the parameters of the image they wish to see based on things like gender, hair colour and ethnicity.

All of this is bad news for content creators. I’d like to see an AI tax to mitigate that kind of financial turmoil. Here is how it would work: if a company uses AI for a commercial process, such as creating a product that someone else may buy, it would be subject to an AI tax. That money would go to a fund used to compensate people, like artists, who lost their jobs because of AI. Maybe that sounds like an overly socialist idea, but it is rooted in my belief that humans should be compensated anytime a robot takes their job. Whether or not a government would enact this is impossible to say. There will be huge pushback from industries who see robots and AI as a way to lessen their labour costs.

Anyway, I’m less worried that AI will bring about massive job loss in the world of art than I am about how it will further complicate its ongoing copyright issues. The compensation and copyright systems in music are already in disrepair. Our current legal system has no clarity on any of this. Recently, an anonymous producer trained an AI on songs from Drake and the Weeknd and made a new track called “Heart on My Sleeve,” using their voices and styles. The song received millions of streams. When people sample someone else’s voice or melody, do they have to pay? When does borrowing become copying? Using artists’ likenesses without credit, in my opinion, crosses a moral line, but it’s murky whether it crosses a legal one. Copyright laws are slippery: in Canada, the creator of the work owns the copyright and can sue someone for stealing it, but we don’t know what happens when a machine recreates it.

Situations like these will become more common across disciplines and risk leaving the entertainment industries in a huge mess. Imagine this: a studio wants to cast Sean Penn in a movie, but he doesn’t want to do it. So, instead, the studio makes a 3D version of Penn using images that already exist online, trains an AI on his voice, then hires a body double and maps Penn’s face on it. Is that wrong? Morally, I say yes. Legally, I don’t know. The technology already exists: the new Indiana Jones movie, which came out this year, de-aged Harrison Ford’s face by using machine learning trained on footage from previous films in the series. We will probably begin to see more cases where people use the faces and voices of deceased actors for their movies. Even weirder, we could also see the creation of composite AIs based on many of the best actors in history. Imagine creating a new movie star with one actor’s eyes, another’s body and a third’s mannerisms. How can human stars compete?

I expect AI technology, as it becomes more sophisticated, to fuel more property battles between studios and artists—it’s already happening in the Hollywood writers’ strike. Actors could add clauses in their contracts that read something like, my face is mine, and you cannot use it anywhere else. This scenario could also play out in the opposite direction. A studio or record label could prepare a contract with a clause like, if we sign you, we own you and your style. People who are desperate to enter the industry might sign something like that. Later, this will lead to a lot of conflict. Under the law, it will be tricky to side with someone who has signed a contract.

This is why we have to agree on good copyright law. I think our faces and voices should be our property. This would present issues for photographers and filmmakers, but surely we can construct a law that allows me to appear in photographs in artistic settings while also allowing me to charge for the use of my own image in commercial settings.

If we do not hash out these questions right now, throats will be cut and artists will no longer be able to make a living. Like Photoshop and Auto-Tune before it, AI is a wonderful tool of inspiration that artists can put in their toolkit. But we should not send this technology into the wild without constraints. If we do not create clear laws now, our arts industries could fall apart.


We reached out to Canada’s top AI thinkers in fields like ethics, health and computer science and asked them to predict where AI will take us in the coming years, for better or worse. The results may sound like science fiction—but they’re coming at you sooner than you think. To stay ahead of it all, read the other essays that make up our AI cover story, which was published in the November 2023 issue of Maclean’s. Subscribe now.