Synthetic stardom: When AI came for Hollywood

In the immortal words of Emily Blunt, “Good Lord, we’re screwed.”

She was on a podcast with Variety Monday when she was handed a headline about cinema’s latest sensation, Tilly Norwood.

Agents are circling the hot property — a fresh-faced young British brunette attracting global attention.

Norwood is AI, and Blunt is PO’d. In fact, she says, she’s terrified.

Told that Tilly’s creator, Eline Van der Velden, a Dutch former actress with a master’s in physics, wants her to be the next Scarlett Johansson, Blunt protested, “But we have Scarlett Johansson.” (Cue the Invasion of the Body Snatchers music.)

Across Hollywood, actresses are cursing Tilly, her Pygmalion Van der Velden, and the increasingly withdrawn men who prefer eternally youthful, preternaturally gorgeous AI replicas. (No Botox or Ozempic required.)

Meanwhile, studio executives are licking their chops at the prospect of more malleable actors. “She’s not going to talk back,” one top talent wrangler said dryly.

They may be alarmed by one AI actress now, but as AI expert Nate Soares, coauthor of the bestseller If Anyone Builds It, Everyone Dies, explains, “AI is less like one actress and more like a puppeteer behind lots of different characters.”

Checking out Tilly’s image, Blunt was clearly unnerved. “That is really, really scary,” she told Variety. “Come on, agencies, don’t do that. Please stop taking away our human connection.”

I fear it’s too late.

Human connection has been eroding for some time. We all live in the Uncanny Valley now, staring into our screens, unsure what’s real or fake, to the detriment of talking, dating, reading, and living.

And we’re getting another jolt about how fast AI is advancing. Just this week, we were inundated with racist, juvenile AI videos posted by Donald Trump, mocking Democratic leaders as the government shut down. The president is wallowing in AI slop.

Sam Altman, the head of OpenAI, debuted Sora — an app that creates alarmingly realistic fake videos. It could be TikTok with more disinformation. A text prompt can conjure terrorist attacks, election fraud, mass protests, war scenes, even disturbing sexual scenarios.

“Increasingly realistic videos are more likely to lead to consequences in the real world by exacerbating conflicts, defrauding consumers, swinging elections or framing people for crimes they did not commit,” The New York Times reported.

The app will further erode truth and trust — in a country where the president promotes false narratives and nearly two-thirds of voters believe we’re too politically divided to solve our problems.

Sora will also let people dismiss real footage as fake. “Until recently,” The Times noted, “videos were reasonably reliable as evidence of actual events. Sora’s high-quality video raises the risk that viewers will lose all trust in what they see.”

Though many in Tinseltown are upset by Tilly and Sora, AI will likely make swift inroads in a degraded Hollywood. The era of blazing movie stars and prestige mass-appeal films is largely gone. (Sydney Sweeney already looks suspiciously like AI.) Now it’s all Marvel, sequels, adaptations, and streaming shows that feel as if they were written by algorithms for consumption while scrolling on another screen.

“I get it, even though I don’t like it,” said Lola Kirke, actress and author of Wild West Village, a collection of essays about New York and her eccentric, creative family. “It’s a business, and they have to keep up with public demand. People are more used to watching face-tuned influencers lip-sync Real Housewives sound bites for 15 seconds than actors telling stories over three acts. Maybe, in some weird way, it will revitalise interest in film and TV? That’s me being optimistic — albeit in a sad way.”

The less optimistic view comes from Jaron Lanier, a top scientist at Microsoft.

He said a studio chief was bragging that AI would save him from paying “all these idiot producers and actors and lighting people and composers and writers and agents.” Lanier told him studio chiefs would soon be expendable too, since everyone will serve at the mercy of “the big computer server at the centre, and Silicon Valley will just roll right over you.”

While Lanier thinks a simulated character here and there is fine, he says it’s urgent to draw a clear line between “AI-generated stuff and reality-generated stuff — to have a system in which we know what’s real and what’s fake.”

“The problem,” he said, “is if you make the whole world run by fakes and simulations, everybody becomes increasingly dysfunctional. People grow alienated, nervous, and unsure of their own value, and the whole thing falls apart. At some point, it’s civilizational and species collapse.”

That, readers, would be less than ideal.

@The New York Times