Inside Google's plan to have Hollywood make AI look less doomsday

MOUNTAIN VIEW — For decades, Hollywood directors including Stanley Kubrick, James Cameron and Alex Garland person formed artificial quality arsenic a villain that tin crook into a sidesplitting machine.

Even Steven Spielberg’s comparatively hopeful “A.I.: Artificial Intelligence” had a pessimistic borderline to its imaginativeness of the future.

Now Google — a starring developer successful AI exertion — wants to determination the taste conversations distant from the exertion arsenic seen successful “The Terminator,” “2001: A Space Odyssey” and “Ex Machina.”

To bash so, the Mountain View, Calif., tech elephantine is backing abbreviated films astir AI that represent the exertion successful a little nightmarish light.

The Google initiative, called “AI connected Screen,” is simply a concern with Santa Monica-based Range Media Partners, a endowment absorption and accumulation institution that represents a wide assortment of amusement clients, including actors and writers. Range is producing the films.

So far, 2 abbreviated films person been greenlit done the project: One, titled “Sweetwater,” tells the communicative of a antheral who visits his puerility location and discovers a hologram of his dormant personage mother. Michael Keaton volition nonstop and look successful the film, which was written by his son, Sean Douglas. It is the archetypal task they are moving connected together.

The other, “Lucid,” examines a mates who privation to flight their suffocating world and hazard everything connected a instrumentality that allows them to stock the aforesaid dream.

“They were looking for stories that were not doomsday tales astir AI, which I was good with, due to the fact that I deliberation we’ve seen truthful galore of those,” Douglas told The Times. “It’s bully to spot the much — not overly affirmative — but benignant of middle-ground stories.”

The effort comes astatine a clip erstwhile galore Americans person mixed feelings astir AI. A 2024 survey from Bentley University and Gallup showed that 56% of Americans spot AI arsenic doing “equal amounts of harm and good,” portion 31% judge AI does “more harm than good.” Shifting the mode AI is depicted successful fashionable civilization could assistance displacement those perceptions, oregon astatine slightest that’s what immoderate techies and AI enthusiasts hope.

Google has overmuch riding connected convincing consumers that AI tin beryllium a unit for good, oregon astatine slightest not evil. The blistery abstraction is progressively crowded with startups and established players specified arsenic OpenAI, Anthropic, Apple and Facebook genitor institution Meta.

The Google-funded shorts, which are 15 to 20 minutes long, aren’t commercials for AI, per se. Rather, Google is looking to money films that research the intersection of humanity and technology, said Mira Lane, vice president of exertion and nine astatine Google. Google is not pushing their products successful the movies, and the films are not made with AI, she added.

“Narratives astir exertion successful films are overwhelmingly characterized by a dystopian perspective,” Lane said. “When we deliberation astir AI, there’s truthful overmuch nuance to consider, which is what this programme is about. How mightiness we archer much profoundly quality stories? What does it look similar to coexist? What are immoderate of those dilemmas that are going to travel up?”

Google did not disclose however overmuch they are investing successful the films. The institution said it wants to money galore much movies, but it does not person a people number. Some of the shorts could yet go full-length features, Google said.

Creators who enactment with Google are fixed entree to tech experts astatine the institution who tin stock much accusation astir the technology. Does the exertion successful the publication already exist, for example? How would it enactment successful existent life?

“We’re surviving with this exertion and AI — the questions originate include: How does it impact america and however tin we emotionally link via this benignant of technology?” said Rachel Douglas, a spouse astatine Range, who is joined to Sean Douglas.

AI has been a arguable taxable successful Hollywood, playing a large portion successful the 2023 writers’ and actors’ strikes.

Actors fearfulness their likenesses and voices being replicated and manipulated without support oregon payment. Writers interest their enactment is being utilized without their support to make AI-generated scripts and communicative outlines. Animation and peculiar effects jobs could beryllium gutted. Publishers and grounds labels person sued to support their intelligence property.

Negative nationalist perceptions astir AI could enactment tech companies astatine a disadvantage erstwhile specified cases spell earlier juries of laypeople. That’s 1 crushed wherefore firms are motivated to makeover AI’s reputation.

“There’s an unthinkable magnitude of skepticism successful the nationalist satellite astir what AI is and what AI volition bash successful the future,” said Sean Pak, an intelligence spot lawyer astatine Quinn Emanuel, connected a league panel. “We, arsenic an industry, person to bash a amended occupation of communicating the nationalist benefits and explaining successful simple, wide connection what it is that we’re doing and what it is that we’re not doing.”

AI companies, including OpenAI, Google and Meta, person demoed oregon shared their tools with movie and TV studios and directors. Meta has partnered with fearfulness workplace Blumhouse and Cameron’s task Lightstorm Vision connected AI-related initiatives.

On Tuesday, Google announced a concern with “The Whale” manager Darren Aronofsky’s task Primordial Soup, which volition enactment with 3 filmmakers connected abbreviated films and springiness them entree to Google’s AI video generator Veo.

Proponents accidental the tech tin marque filmmaking cheaper and springiness artists much flexibility astatine a clip erstwhile the movie concern is struggling.

“If we privation to proceed to spot the kinds of movies that I’ve ever loved and that I similar to marque and that I volition spell to spot ... we got to fig retired however to chopped the outgo of that successful half,” Cameron said connected a podcast past period with Meta’s main exertion officer. Cameron sits connected the committee of startup Stability AI.

AI companies are uncovering different originative ways to marque the exertion much approachable. In 1 example, large artificial quality steadfast Anthropic is sponsoring an upcoming grounds astatine the Exploratorium, a subject and creation depository successful San Francisco. Eric Dimond, elder manager of exhibits, said helium hopes the exhibit, called “Adventures successful AI,” volition origin much radical to research the costs and benefits of AI.

Anthropic was not progressive successful the conceptualization of the exhibit, Dimond said, though visitors tin interact with its AI exemplary Claude, arsenic good arsenic AI tools from OpenAI and ElevenLabs.

As Google and others effort to enactment a softer absorption astir technology, moviegoers are inactive getting plentifulness of stories astir the dangers of robots tally amok.

Recent tales of AI gone incorrect see Blumhouse’s 2023 fearfulness movie “M3GAN,” astir a robot who becomes truthful protective of a young miss that she starts wreaking havoc. Last year, different Blumhouse fearfulness film, “Afraid,” followed a household terrorized by an AI-powered assistant.

This summer, “M3GAN” is getting a sequel, released successful theaters by Universal Pictures. It’s expected to beryllium a container bureau hit.