“In this business, you’re manufacturing and distributing and selling brain states,” says Yves Bergquist, the director of artificial intelligence and neuroscience in media at the Entertainment Technology Center, Hollywood’s low-profile, semi-secretive think tank. It operates out of USC’s School of Cinematic Arts.
Since 2016, Bergquist has been conducting a series of artificial intelligence research projects in coordination with all the major Hollywood studios, which fund the ETC (platinum level sponsors give at least $100,000 annually). One study uses AI to scrutinize how TV and film fans spread information on the internet. Another looks into how character development and certain plot points may affect box office returns, depending on genre. A third number-crunches large-scale data to make better scheduling decisions. And a fourth endeavors to learn more about Chinese film audiences by tracking and analyzing film-related online conversations in the country.
“One of the heads of a really large studio told me, ‘I want to take opinion out of development,’ ” says Bergquist at the ETC offices in September. “Right now, you have a bunch of people, mostly older white people, just using their intuition to decide, ‘Hey, does this make sense for us to develop?’ So what we’re doing here is we’re developing a language that is based on data that both the creatives and the executives can speak together.”
Bergquist, a 47-year-old Frenchman, presents as the market-disrupting sage — flowing hair, an open-collared dress shirt, a thick metal bracelet featuring a coiling snake. Online, he tweets as @punkstrategy and refers to himself as a “Machine Intelligence Warlord.”
For many in the industry, Hollywood’s brewing AI revolution has become an intriguing and anxious topic, defined on one end by Netflix’s opaque audience data and on the other by producer Ryan Kavanaugh’s notorious Relativity Media, which touted a magic algorithm for box office hits before its implosion. But Bergquist portrays himself as an incrementalist. He’s quick to note that he’s not in the business of algorithmic predictions, which he believes are responsible for the snake-oil reputation of his specialty in Hollywood. “The notion that AI is going to write a [hit] script is just complete garbage,” he says, explaining that there are just too many variables to account for. “It’s a black box.” He sees his work as being “very early on a 20-year journey” and prefers to cite modest takeaways: that audiences favor linear-thinking protagonists, for example, and that they prefer narrative problems solved by a group of people rather than by an individual.
Bergquist believes that his research uncovering patterns in narrative data will be used to optimize studio marketing strategies and to decrease development costs. One seasoned TV executive told the ETC that if network pilot season could be made even slightly more economical by AI, tens of millions could be saved each year. “You don’t have to be perfect,” said the exec. “Just make me a little smarter around the edges and I can make better decisions.”
While Bergquist acknowledges “a lot of push-back” from talent wary of standardization, he foresees how “computational methods” will actually help identify stale storytelling: “We’ll be able to say, ‘Well, this character journey in the female lead is very traditional in the horror genre,’ ” he observes.
Bergquist — who also runs what he terms a “stealth” Hollywood AI startup called Corto — declines to share his research beyond a few slides he occasionally deploys on the Big Data presentation circuit, citing confidentiality restrictions and the work’s still-germinal nature. But as The Hollywood Reporter has learned, Bergquist’s tight-lipped R&D in the space is only the second-most-intriguing thing about him: Yves Bergquist was once known as Alexis Debat, an on-the-rise national security scholar in Washington, D.C. — until he left town in 2007 after being exposed for an array of intellectual fabrications. None of his studio financial backers have been made aware of this history, including Disney, whose sister division ABC News once employed him as an on-air expert before he was forced to resign.
“Should I have blown my brains out?” Bergquist says when asked about his secret past. “Was that the right thing to do? Or was the right thing to do to move forward and become a better person and try to rebuild my life?”
In an age of cancel culture and attempted comebacks, the Bergquist/Debat saga is an allegory of the possibilities and perils of forgiveness. It’s also a reminder that the path to professional clemency is often murky. Bergquist says that while he’s been living under a new name, he’s certainly still haunted by what he’s done. “It’s very much a recovery process,” he explains. “This thing is very present. It’s not forgotten or forgiven.”
Bergquist explains that he’d long anticipated being discovered. “I was waiting for this moment,” he says. “I don’t seek absolution. I seek a second chance.”
In the years after Sept. 11, as America became enmeshed in Middle East conflict, Alexis Debat emerged as a war expert in Washington, D.C. Then in his 30s, Debat, who told associates he’d previously worked as an official for the French Ministry of Defense, directed the terrorism and national security program at the Center for the National Interest, a conservative think tank set up by Richard Nixon after his presidency. He also authored articles with a realpolitik bent — including one titled “In Praise of Warlords” — for the affiliated foreign policy journal The National Interest, whose honorary chairman is Henry Kissinger. Debat’s highest-profile gig, though, was as a frequent on-air consultant for ABC News, where starting after the Sept. 11 attacks he provided commentary and contributed to War on Terror-related reporting, including by its chief investigative correspondent, Brian Ross.
Then, in 2007, Debat’s professional life imploded. He was exposed for falsifying his résumé with such fictional entries as a Ph.D. from the Sorbonne. “I didn’t need a Ph.D. for anything that I did,” he says. “So that’s what makes it particularly stupid.” He also was accused of faking a series of interviews with famous figures including Kofi Annan, Michael Bloomberg, Bill Gates, Alan Greenspan and then-presidential candidate Barack Obama. In the aftermath, Bergquist claimed that he had subcontracted out the work to a third party and was himself duped, though follow-up inquiries failed to corroborate this.
According to a 2007 investigation on Debat in Mother Jones, his work at ABC led the network to report that the U.S. had been supporting an Iranian ethnic minority group to attack the Tehran regime. But subsequent reporting didn’t bear this out: The leader of the group objected to ABC’s story and the Iranian foreign ministry lodged a complaint with the U.N. asserting that the TV report was proof that the U.S. was attempting to destabilize its government. Bergquist stands by this reporting: “Multiple sources had given me an extreme amount of detail about this operation.”
Media covering l’affaire Debat described him as a “fraud” (Talking Points Memo) and “flimflam man” (NPR). Some observers wondered whether he was a lone-wolf fabulist, or an instrument of disinformation for more shadowy figures. While Bergquist now acknowledges his “mistakes,” he contends the armchair psychologizing of his behavior was out of line.
“A lot of the commentary was, ‘He’s a con man,’ ” he says, explaining that the description is unfair. “It’s to say, ‘This guy is constitutionally a liar and constitutionally a fraud’ versus ‘This guy did some really sleazy stuff.’ “
Bergquist points to the conclusion of ABC’s probe of his work. David Westin, then the president of its news division, issued a memo noting that “none of these discrepancies would rise to the level of a formal, on-air retraction because none of them was material to the substance of our report.” Critics countered that the network’s self-inquiry had been a sham. To this, Bergquist scoffs, pointing to ABC’s axing of Ross in 2018 after issues arose with his former boss’ work.
Regardless, the episode shone a harsh light on the common TV news practice of employing ideological partisans for on-air commentary and (often uncredited) back-channel reporting. Debat soon lost all his gigs and disappeared, not just from his appearances on Jim Lehrer’s NewsHour and C-SPAN, but from D.C.
“What happened to me was horrific,” says Bergquist. “I lost everything. But it gave me a phenomenal opportunity to be reborn. That guy was left behind.”
Shortly after Bergquist’s public collapse, he met the woman who’d eventually become his wife. She had a 5-year-old daughter. “You can imagine becoming a father in the worst recession since the ’30s, with my Google results,” he says. “Those were very difficult years.” The new family moved to California, where he started going by his middle name, Yves, and went back to school, receiving an MBA from what is now the Middlebury Institute of International Studies at Monterey. While there, he learned of the nearby Singularity University, the futurist Ray Kurzweil’s then-nascent graduate program, where he was brought on as a teaching fellow and was first exposed to the frontiers of artificial intelligence.
By the account of Bergquist — who claims he’s spent years in therapy “unpacking” his transgressions, believing they stem from the insecurity of a fresh émigré from France looking for an edge in the nation’s careerist capital — AI appealed to him for its contrast to his prior profession. “It’s an area where you’re accountable to specific metrics,” he says. “In my previous life, you’re accountable to opinion. ‘I think that al-Qaida is this, so we should fight it like that.’ With machine learning, you are accountable to your models. They either work or they don’t. You can’t bullshit your way out of it.”
Bergquist learned that his past would follow him. “I didn’t start out telling people,” he says, explaining that after he was quickly promoted at Singularity (where he was still using his surname, Debat), someone wrote an anonymous letter to his superiors, “and they fired me. My son was a week old.” At that point he took his wife’s surname and, he says, resolved to be more up-front about his checkered history. After a period in the consulting wilderness, he was hired at L.A.-based pop culture polling website Ranker. Its CEO, Clark Benson, confirms that Bergquist was suitably candid about his past.
The current director of the ETC, Ken Williams, met Bergquist in 2014, when the think tank was working with raw data provided from Ranker. Bergquist later pitched Williams on a project of his own, to explore the ways in which Hollywood could benefit from the new computational methodologies known in the AI community as machine learning. Soon, with the ETC board’s approval, Bergquist began as a pro bono consultant, running the center’s newly minted data and analytics project. Bergquist didn’t then tell Williams about his past. “If he had come in off the street and that was on his résumé,” acknowledges Williams, “I can’t say that I would have been quite so supportive.”
The ETC maintains a symbiotic relationship with Hollywood akin to the close ties between the MIT Media Lab and Silicon Valley, or RAND and the defense sector. Spurred into creation by USC alumnus and ILM founder George Lucas in 1993, it benefits from six-figure annual financial support from each of the major studios, whose chief technology officers and senior tech execs sit on its board and participate in its working groups. Over the years, its tests, models, whitepapers and invite-only conferences have served as the intellectual apparatus for everything from the industry’s move to digitalization to its embrace of standardized cloud systems. Recently, the center has been developing new internal research focused on applications for blockchain in entertainment.
Since joining the center, Bergquist also has launched an under-the-radar tech startup, Corto, which has its own reciprocal relationship with the ETC: The fledgling firm, supported by interns from USC, builds software informed by his directed research work there.
Corto’s clients license its products. Bergquist won’t specify those clients’ names, except to say that they include some of the sponsors of the think tank. He notes that the firm is still in “early-stage mode” and self-financed. “Corto and ETC have discussed the possibility of some type of equity grant or licensing arrangement should any of the ETC research be commercialized,” Bergquist explains.
Williams, who says he has no involvement in Corto, stresses that the relationship between the startup and the think tank is copacetic. “None of our lead consultants are exclusive to the ETC,” he says. “They are all expected to be independent consultants and practitioners, much the way some of the valued professors at [USC’s] film school are actually still working artists and technicians. So our structure has always been that we want working practitioners who are active in the field.” (Williams is best known as the co-founder of Sony Pictures Imageworks, the studio’s VFX and computer animation unit.)
In November 2017, more than a year and a half after the ETC board approved his program, Bergquist approached his boss about his past. “I wanted to tell him for a very long time,” says Bergquist. “I was very worried that someone would come calling before I told him. It’s just not a fun conversation. I was really afraid he’d say, ‘We love you but it’s not something we can live with and we need to part ways.’ I was very scared that he’d have to let me go.”
Bergquist concedes he purposefully waited those many months to tell Williams about his secret. “I wanted the time to build a successful track record to weigh that against my past,” he says. “These are hard, messy decisions.”
For years, Williams chose to keep his dozen-member board out of the loop about Bergquist’s past. “Frankly, I thought his work would be compromised,” he says. His stance was defined by how “very happy” he was with Bergquist’s work. “I think people deserve a second chance if they’ve earned it, and I thought he earned it,” he says. In May 2018 he furthered his approval by putting Bergquist on the center’s payroll.
In recent years, multiple board seats have been held by executives at subsidiaries of Disney. Now the media company was unknowingly funding Bergquist again. Asked about this, Williams says, “It wasn’t something I gave a lot of consideration to.” Disney didn’t return a request for comment.
This June, as Bergquist’s profile continued to rise, and he became a frequent presence on the tech conference circuit, Williams did confide in one board member, USC School of Cinematic Arts dean Elizabeth Daley. “I started to realize,” says Williams, “that this might be something that could bubble up.” Daley tells THR she didn’t seek further input from her board colleagues, deferring to Williams’ wisdom. “It never seemed significant enough,” she says of the situation. “I would hardly go around the CEO to talk to the board.”
By the time Daley learned about Bergquist’s past, USC was already contending with headline-making ethical crises at its medical and business schools, as well as its outsized role in the college admissions scandal. Daley says that she, too, was sympathetic to the idea of a second chance for Bergquist and that, as far as she knows, “we have 12 years of a pristine record on this guy.”
When asked whether the ETC, as an affiliate of a top research university, should be employing someone who bears a public record of falsehood, Williams and Daley both note that Bergquist isn’t a teacher. A School of Cinematic Arts spokeswoman later underscores the point: “Per Ken Williams, Yves never represented himself as an academic researcher — rather as an industry researcher and consultant.”
Bergquist contends that his dishonesty during his Debat days makes him more trustworthy now. “Having that in my past creates a higher level of accountability because you have to work harder to be on par with the value of someone who has not had that,” he says. “So I put a lot of pressure on myself to deliver outstanding work — even if the people I’m doing that work for don’t know about that past.”
Bergquist believes in himself, along with the people whose faith he’s secured. “I presented the research agenda to the board in 2016 and they thought there was enough value to let me do what I do,” he says. “They aren’t idiots. They’re extremely smart senior executives at big companies. If I was an impostor or I was not good at what I did, I wouldn’t have lasted as long as I have.”
On Nov. 8, Bergquist spoke in Hollywood at the Infinity Festival, a conclave of tech and entertainment “thought leaders,” who gave future-looking presentations like “Studio in the Cloud,” by an Amazon executive, and “Fandom Across Life Stages,” with a Warner Bros. consumer insights vp.
Bergquist’s lecture sought to dispel the notion that data wonks and industry creatives are inherently at odds. “If you take a class in writing at a film school, on the first day you’ll hear that stories are a set of steps to solve a problem,” he said to the crowd, itself oblivious to his story and the problems it conjured. “What else is a set of steps to solve a problem?” he posited. “Algorithms.”
This story first appeared in the Nov. 13 issue of The Hollywood Reporter magazine. To receive the magazine, click here to subscribe.