This is the web version of dot.LA’s daily newsletter. Sign up to get the latest news on Southern California’s tech, startup and venture capital scene.
Disney’s Research Studios, headquartered locally in Glendale, California, houses the teams working on the company’s most cutting-edge scientific and technological innovations. Recent announcements included the company’s newly-patented “robot sherpas,” which will assist guests while navigating themselves throughout the company’s chaotic and crowded theme parks, as well as a new twist on Augmented Reality called “PuppetPhone,” allowing users to manipulate animated characters using handheld devices.
The company’s latest project – introduced this week via a new demo video and whitepaper – uses artificial intelligence and machine learning to automate the digital aging and de-aging process, all without the intervention of an animator or special effects wizard. (In other words, after an actor’s face is filmed and input into the system, the new tools can automatically make the performer appear older or younger, as Disney previously did with actor Mark Hamill in TV series “The Mandalorian.”)
To achieve the dramatic effect, Disney researchers first created an original database including thousands of synthetic human faces. These automatically-generated faces were then aged and de-aged with existing machine learning tools, and the results of this process were used to train a brand new neural network, dubbed the Face Re-Aging Network, or FRAN.
When a new headshot is fed into FRAN, the software first predicts which facial features would be most impacted by the aging process, such as wrinkles and skin texture. Then, results are layered over the original photograph, preserving the actor’s appearance even if they move their head or the lighting shifts within the frame.. Perhaps most importantly on a practical level, FRAN is not a closed system; it can interact with a team of animators and artists, and could theoretically be integrated into an existing film or show’s production process.
Films and TV shows have routinely used computer animation to artificially make actors appear older or younger, allowing them to portray characters across a wider range of ages and appearances, for well over a decade now. The technique made its formal debut in 2006’s “X-Men: The Last Stand,” when younger versions of Patrick Stewart’s Professor X and Ian McKellan’s Magneto appeared in a flashback, and received a lot of attention for its use throughout David Fincher’s 2008 film “The Curious Case of Benjamin Button.” More recently, Lucasfilm de-aged Harrison Ford for a flashback scene in the new “Indiana Jones” film.
Still, the process remains divisive among film fans, even when adopted by widely-respected filmmakers. Martin Scorsese got a lot of side eyes and muted responses when he digitally de-aged Robert De Niro in “The Irishman.” That film also suffers from some visual hiccups that Disney’s new AI tool can’t solve; De Niro’s face may look smoother and less wrinkly, but the actor still physically moves around like an older man, enough for viewers to notice with the naked eye.
Automating these kinds of repetitive tasks that were traditionally the domain of salaried employees obviously poses a tremendous cost-saving benefit to The Walt Disney Company, and it has been a priority for some time. In 2020, Disney’s same Research Group developed a similar algorithm for so-called “face swaps,” producing high-quality enough results to integrate them into an actual studio film.
Disney already massively increased its reliance on visual effects artists as a cost-saving measure over physical production crews, most of whom are unionized and therefore more expensive to hire than teams of animators. That brought on its own challenges, as the increasingly intense demands of major content producers – including Disney – has put a tremendous strain on visual effects houses and their staffers. In the most famous example of the uncertainty and chaos plaguing the industry, in 2013, visual effects studio Rhythm and Hues won the Oscar for their effects work on the film “Life of Pi” and then declared bankruptcy just 11 days later. (“Life of Pi” was produced by 20th Century Fox, which has since been acquired by… you guessed it.)
So while additional savings on movie and TV visuals is certain to please the company’s board, not to mention shareholders, it’s the latest setback for Hollywood animators, a group that’s already contending with a number of challenges, including key negotiations with the studios that employ them, not to mention a potentially softening audience for their actual films in theaters. - Lon Harris
An LA Agency Allegedly Scammed Influencers Out of Brand Deals
In total, The Carter Agency’s actions have affected more than 130 influencers, including those signed to Ben Popkin’s company and those who say he falsely claimed to represent them.
Was Neuralink’s Demonstration Just One Big Recruiting Drive?
Elon Musk’s controversial, experimental brain computing company Neuralink held a live demonstration last night, aimed at proving to the world that its technology is safe and close to viable human trials.
Snap’s RTO Plan Is Meant To Boost Productivity. It Could Do the Opposite
Snap is the latest major tech company to bring the hammer down on remote work: CEO Evan Spiegel told employees this week that they will be expected to work from the office 80% of the time starting in February.
Get Tickets to the #LongLA Holiday Party with dot.LA
Join dot.LA and other great organizations for the #LongLA holiday party on Thursday 12/15! More information and tickets available here.
What We’re Reading...
-- At the New York Times' DealBook conference, TikTok CEO Shou Zi Chew discussed moving the platform’s data to a new U.S.-based cloud infrastructure.
-- Wired looked into Amazon’s rapid development of its own air cargo business, Amazon Air, and how it’s helping the company move away from dependence on UPS and FedEx.
-- Also at DealBook, Netflix co-CEO Reed Hastings conceded he’d been “wrong” about bringing advertising to the platform.
-- YouTube launched an official podcast, “Like & Describe,” tracking the platform’s most popular trends, hosted by creator MatPat.
--
How Are We Doing? We're working to make the newsletter more informative, with deeper analysis and more news about L.A.'s tech and startup scene. Let us know what you think in our survey, or email us!
- Art Created By Artificial Intelligence Can’t Be Copyrighted, US Agency Rules ›
- LA Is Emerging As an 'Early Adopter' of Artificial Intelligence, Brookings Finds ›
- A Decentralized Disney Is Coming. Meet the Artists Using AI to Dethrone Hollywood ›
- Anime Artists Are Mad Netflix Japan Used 'Image Generation' - dot.LA ›