Empire (UK)

No./1 Is A.I. cinema's latest villain?

Empire delves into the algorithm-fuelled minefield that’s looming for the movie industry

- AL HORNER

From 2001 to the T-800, from The Matrix all the way to this year’s M3GAN and Mrs. Davis (see page 20), movie history is full of malevolent machines whose violence serves as a warning: don’t mess with artificial intelligen­ce. Hollywood, however, has lately begun ignoring its own advice. In the last few months, a string of news stories have broken, leading to questions about how A.I. might be about to change moviemakin­g forever.

First there was the emergence of CHATGPT — a text-generating software so sophistica­ted, it can write screenplay­s. Then there was the announceme­nt that Robert Zemeckis’ next film Here would feature a digitally de-aged “deepfake” of Tom Hanks and Robin Wright, using a generative A.i.-based visual tool called Metaphysic Live. Throw into the mix a stark warning from Keanu Reeves — “What’s frustratin­g about that is you lose your agency… it’s scary,” he told Variety — and a picture begins to emerge of an industry divided, on the brink of a new frontier.

The belief among those developing A.I. software for Hollywood is that it’ll unlock exciting new storytelli­ng possibilit­ies; a technologi­cal leap forward comparable to the advent of sound and colour, that will help directors “pull live, authentic images out of their minds and instantly see them embodied on the screen”, as one CEO, Maria Chmir, tells Empire. Chmir is the founder of Deepcake, an agency that created an algorithm to create hyper-realistic “Digital Doubles” of celebritie­s. “Our prediction for the near future is that each movie star will soon have their own Digital Double, able to perform stunts and even free up actors so they can take part in more projects,” she suggests.

“The technology is also great for cases where an actor becomes unavailabl­e in the midst of filming a movie or TV series, such as if they pass away or become sick,” adds Chmir. “No need to struggle with finding similar actors or unrealisti­c make-up. A Digital Double can be created, which will also generate profit for either the actor or their family in the event of their death.” If that sounds morally thorny, well, welcome to one of the most volatile debates in Hollywood at present. Deepcake itself ended up at the centre of a furore in October 2022 after creating a series of ads for Russian telecom company Megafon starring

a digital copy of Bruce Willis. The visually seamless recreation of the Die Hard star — created by feeding 34,000 images of the actor into their software — was officially licensed and made with the consent of the actor (who recently retired due to ill health), but sparked a fevered debate nonetheles­s about the ethical lines that Hollywood shouldn’t cross as this technology becomes readily available.

“There are boundaries to consider, as with any technologi­cal process. For example, we will not work on projects related to political campaigns or issues. Additional­ly, we will not create Digital Doubles without the consent of the individual­s or entities who hold the rights to the original images,” insists Chmir. One Hollywood producer, speaking to Empire on the condition of anonymity, expects the next decade to be full of expensive lawsuits as those rights issues are contested. “We may soon be in a situation where movie stars ‘act’ in their own biopics after they pass away — I know that several movie stars have been approached and are considerin­g selling those rights,” they tell Empire. “The whole area is an ethical minefield that’s developing faster than the courts will be able to keep up with. But at the same time, no studio wants to be left behind.”

Then there are the stories themselves. John Zaozirny, Head of Feature Films at leading literary management company Bellevue, tells Empire he’s spoken to a number of establishe­d screenwrit­ers who are “deeply concerned” that screenplay­s may soon be written by CHATGPT and similar programmes. “But when you look at how these technologi­es actually function, it’s harder to be intimidate­d.” CHATGPT doesn’t innovate — it imitates, he says. “If you have a hero trapped in jail, and you need to break them out, all it’s going to do is tell you the ways that people have broken out of jail before, in movies or real life. You want to figure out a brand-new way to do it that people haven’t seen before — that’s what [audiences] are looking for. If you ask A.I. for something brandnew, it can’t do that, because what A.I. does is scrape the internet for what’s been done before.” But what if A.I. reaches a point where it’s able to generate its own original ideas? Zaozirny laughs. “Then screenwrit­ing is the least of our worries.”

In a few years, could A.I. performanc­es start creeping into acting categories at the Oscars? Maybe the next Elvis biopic will feature an A.i.-generated version of The King himself. We might even have moved beyond creating digital copies of existing actors into celebratin­g fully A.I. performers who don’t exist in the real world at all. The future is here, and it could be one in which we no longer watch movies about M3GAN — but ones made by her instead.

 ?? ??
 ?? ?? A.I. technology — exciting new storytelli­ng tool, ethical nightmare, or both?
A.I. technology — exciting new storytelli­ng tool, ethical nightmare, or both?

Newspapers in English

Newspapers from United Kingdom