Note: This article is based on Variety Intelligence Platform’s special report “Generative AI & Entertainment,” available only to subscribers.
Advances in deepfake technology raise the prospect that the day will soon come when the digital likeness of actors can be brought to life for movies and TV without their needing to even lift a finger.
These likenesses have been variously referred to as personal avatars, digital twins or digital doubles. While entirely new synthetic personalities can be created, generative AI can also yield realistic replicas of celebrities.
The prospect of acting avatars was thrust front and center last year when, months after Bruce Willis retired due to illness, he appeared in a commercial for a Russian wireless company thanks to an AI firm, Deepcake, that was able to re-create him. While it was erroneously reported he had sold off his performance rights to Deepcake, it turned out the commercial was more of a one-off usage of the actor’s likeness.
Typically, there are two types of avatar tools: one that generates realistic and reactive moving and speaking avatars given video sample footage, typically a few minutes long, of a specific person; and a second type that generates more constrained moving avatars given a still image that is then manipulated to look responsive (blinking, grinning).
Deepfake video and audio could enable talent and media businesses to produce and scale content output that was previously impossible, whether with existing tools or, frankly, the limits of physical space and time.
Realistic avatars and voice clones can be effectively deployed into new environments or productions without the talent’s physical presence or contribution being required, even in instances beyond an individual’s physical ability — for example, if an actor is already engaged in a different project, is sick or even dead.
That ability to replicate a specific actor for future or additional performances he or she might not otherwise have been able to execute is new. Theoretically, it offers actor talent more commercial opportunities, without doing much or any physical work to appear or perform in new productions. Optimistically, generative AI could have a multiplier effect on talent opportunity if it allows them to accept simultaneous projects, albeit where some use their digital likeness.
In a recent interview with Collider, director Joe Russo recently reflected on future possibilities in which AI would not only resurrect an actor but even bring the viewer into the production.
“You could walk into your house and save the AI on your streaming platform. ‘Hey, I want a movie starring my photoreal avatar and Marilyn Monroe’s photoreal avatar. I want it to be a rom-com because I’ve had a rough day,’ ” Russo theorized. “And it renders a very competent story with dialogue that mimics your voice. It mimics your voice, and suddenly now you have a rom-com starring you that’s 90 minutes long.”
Used correctly, involving AI eclipses limitations that were previously inherent. This could lead to advantageous outcomes including localizing and personalizing content at previously unimaginable — or unmanageable — scale, with some experiments already materializing.
For example, voice clones and visual dubbing techniques can be deployed to produce any number of local-language versions of films or TV; market-specific podcast, radio or TV and video ads could play in local markets; or even high volumes of consumer-specific experiences could air — such as personalized video messages on Cameo — all without requiring talent to physically read or perform each one individually.
Of course, this will require updated thinking to protect talent and ensure fair compensation and clear consented use. At a minimum, talent contracts will need to evolve, though likely legal and regulatory updates will also be needed.
New language has already begun to appear in some actor contracts that includes “simulation rights” to performers’ images, voices and performances, effectively permitting future productions to use their synthetic likeness. This has also raised questions about how an actor, or deceased actor’s estate, can or should be able to give consent for certain uses of their synthetic likeness.
While there is evident opportunity to expand talent opportunities, there would seem to be equal potential for talent exploitation and diminished future opportunities if they unwittingly or desperately sign away or sell off use of their digital likeness absent compensation and consent.
Read more of VIP+’s AI assessments:
• Takeaways for diligence and risk mitigation
• Will gen AI supplant or supplement Hollywood workers?
• Gen AI aims to bring efficiency to the production process
• Streaming audio may never be the same after AI — and that’s good
Plus, dive into the expansive special report …
