Tuesday, November 28, 2023

‘Embrace it or risk obsolescence’: how will AI jobs affect Hollywood?

Must read

Since actors joined writers on the picket line in July, the two guilds, on their first joint strike since 1960, have found a common locus of fear and frustration: the potential encroachment of artificial intelligence on their livelihoods.

The Writers Guild of America (WGA) warns of the possibility that generative AI – the type of machine-learning systems capable of creating text, images and video, such as OpenAI’s ChatGPT – could allow studios, represented by the Alliance of Motion Picture and Television Producers (AMPTP), to cut costs by forgoing the employment of human writers for AI-produced scripts. The Screen Actors Guild (Sag-Aftra) is concerned with the use of digital likenesses, particularly after Duncan Crabtree-Ireland, the guild’s chief negotiator, said studios proposed to pay background actors for a day’s work to use their images in perpetuity. (The AMPTP has disputed this claim as a “mischaracterization”.)

All the while, entertainment companies, or tech conglomerates with entertainment divisions, have continued to expand their human staff tasked with the development, research or management of AI. Last month, Netflix made headlines for a job listing for an AI product manager with an annual salary somewhere between $300,000 and $900,000 (according to Sag-Aftra, 87% of the guild’s actors make less than $26,000 a year). A review of Disney’s job board by the Guardian found at least a dozen roles related to machine learning, several of them within its media & entertainment division. Tech companies such as Amazon and Apple have, of course, numerous open machine learning positions, with some specifically tied to entertainment (a Seattle-based AI role for Prime Video Personalization and Discovery offers, according to the listing, “a once-in-a-lifetime opportunity to shape the future of TV for billions of viewers worldwide. We know our future success is inextricably tied to being a center of excellence in machine learning science and we invest in it.”).

Hollywood’s quiet AI hiring spree, first reported by the Hollywood Reporter and the Los Angeles Times, is not necessarily tied directly to AI-generated scripts or actors’ likenesses, nor are all the positions related to generative AI, the subject of much ethical debate and concern. But taken together, the push to expand AI employment indicates an industry-wide arms race to build up companies’ machine learning capabilities, cutting across many aspects of the business. “All of the studios see the opportunity,” said Dawn Chmielewski, a US entertainment business correspondent at Reuters and the co-author of Binge Times, a book on Hollywood’s streaming wars.

“Obviously, cost is a consideration, but also keeping abreast of the pace of technological change. The studios are not unfamiliar with how Silicon Valley can create a technology that suddenly and dramatically changes their business.”

Some of these positions build on existing AI capabilities within entertainment companies, which already use machine learning for recommendations, advertising and dubbing of foreign languages. Disney is hiring for US-based machine learning engineers for its recommendation algorithms at Hulu and Disney+; Netflix seeks an applied machine learning scientist for the globalization team ($150,000 to $700,000) to “develop algorithms that power high quality localization at scale” and “shape the future of localization and global entertainment at Netflix”.

Sony has open roles for use of AI in video games and machine learning audio processing for, among other things, post production of music and movies, as well as ethics-related roles such as AI ethics technical program manager and AI privacy and security project manager. (All roles are part of Tokyo-based Sony Group, which includes Sony Pictures Entertainment and opened its own AI ethics office in 2021.)

A striking actor outside Netflix’s Los Angeles office. Photograph: Shutterstock

It’s difficult to ascertain the scope of each role, and its potential impact on the entertainment business, from the outside (through a spokesman, Netflix declined to comment this article; Disney, Sony and Amazon could not be reached for comment). But the picture of current investment in AI by entertainment companies is slowly coming into view.

“Generally speaking, there’s a lot of fear of missing out that’s really sweeping the industry,” said Ben Zhao, a professor of computer science at the University of Chicago. Industry executives “are not necessarily sure about what it is that they want” as “companies are still trying to figure out their AI strategy”.

The hiring boom goes beyond entertainment; nationwide, companies are procuring “head of AI” leadership positions with vague mandates in an effort to get ahead of the curve, or out of fear of being left behind; the number of people in AI leadership roles in the US has grown threefold in the past five years, according to data from LinkedIn. “We are all going to be touched by this technology,” said Chmielewski. “And the studios certainly see an imperative to embrace it or risk obsolescence.”

For entertainment companies, the AI hiring boom reflects “a combination of trying to find people with the business acumen to understand what the market needs, what they can do with current AI, and how they can best leverage advancements in AI, particularly generative AI”, said Zhao.

Disney, in particular, is making a concerted push toward AI, framed as part of the company’s long tradition of embracing new technology or as part of an inevitable march of progress. (Disney Studios in Burbank, for example, has listed a job for a senior machine learning engineer to “shape the next generation of creative and production technology and directly help drive innovation across our cinematic pipelines and theatrical experiences”.)

The company recently formed an internal taskforce to study AI and examine how it could be used across divisions, and the chief executive, Bob Iger, has indicated that the company sees AI, including generative AI, as a clear part of its creative future. “Nothing is going to stop technological advancement,” Iger said at a company town hall last November. Generative AI technology, he added, is “something that at some point in the future the company will embrace”.

“In fact, we’re already starting to use AI to create some efficiencies and ultimately to better serve consumers,” Iger said during an earnings call this July, as reported by journalist Lee Fang. Asked about the framing of AI as inevitable, Zhao pointed out that “AI is being touted as an efficiency improver. It’s worthwhile to think about the ultimate result of that efficiency” – namely the prospect of lost human jobs.

Generative AI is, indeed, already shaping the things we watch. It’s used in visual effects, to help with dubbing for Netflix’s many foreign language shows, or to produce a small version of a real rival soccer coach for games in Apple’s Ted Lasso, or to create a full younger version of Will Smith in Paramount’s 2019 film Gemini Man. Disney-owned Marvel Studios controversially used AI to generate the opening credits for its recent series Secret Invasion.

In July, Netflix premiered a new Spanish reality dating series, Deep Fake Love, in which scans of contestants’ faces and bodies are used to create “deepfake” simulations of themselves. The company’s gaming department has used AI to produce narratives and dialogue. In the documentary world, film-makers have used generative AI to recreate the voices of the late Anthony Bourdain and Andy Warhol to bridge archival gaps, and to create “digital veils” – fake faces – to protect interview subjects in Welcome to Chechnya, a documentary on the Russian government’s persecution of LGBTQ+ people.

Annie Murphy in Black Mirror.
Annie Murphy in Black Mirror. Photograph: Netflix

For now, the major fears of writers and actors – the kind depicted in the Black Mirror episode Joan Is Awful, in which a Netflix stand-in streaming service airs an entirely AI-created show – are not yet realized. But from a capability standpoint, “writers and actors could both be replaced by AI in the near future, along with everyone who works to create a movie using real human actors. We could see entire movies generated by AI without any human involvement,” said David Krueger, a researcher with the University of Cambridge’s Machine Learning Group.

“This probably won’t happen in the next few years, but could easily happen within the decade.”

This possibility has underscored actors’ and writers’ demands for protections over AI, from what creative material (performances, text) can be used to train models to protections against AI-generated scripts. The AMPTP has released several public statements on the matter of AI, which “raises hard, important creative and legal questions for everyone” that “requires a lot more discussion, which we’ve committed to doing” and noting that according to the current agreement with the WGA, AI-generated material would not be eligible for credit as “only a ‘person’ can be considered a writer”.

“We are still in the early stages,” said Zhao. “There is so much hype that dominates conversations. And it is very difficult to see facts from the hype” – particularly on the capability and risks of generative AI. It remains unclear what the limits on those will be. But the race for AI in a post-strike Hollywood is already under way.

Latest article