Hollywood celebs are scared of deepfakes. This talent agency will use AI to fight them
Artificial Intelligence,Jobs, Labor & Workplace
Wendy Lee April 2, 2024
As advancements in artificial intelligence proliferate, talent agencies are bulking up their defense
s
to protect Hollywood stars against
misleading, manipulated
images or videos that can put their clients
at riskin compromising positions
.
The rise of generative AI and “deepfakes” or videos and pictures that use a person’s image in a false way have led to the wide proliferation of unauthorized clips that can damage celebrities’ brands and businesses. These clips purport to show famous people saying and doing things they never said or did.
For example:
fake nudes of a famous person, or
videos crafted to make it
look like
a Hollywood star is endorsing a product
they haven’t actually used
.
And the problem is expected to grow. Now there are technological tools that use AI to combat that threat, and the entertainment industry has come knocking.
Talent agency WME
has inked
a partnership with Loti,
a Seattle-based firm that
specializes in software
used to flag
unauthorized content posted of
their
clients’ likenesses on the internet.
The company, which has 25 employees, then quickly sends requests to online platforms to have those infringing photos and videos removed
.
Financial details of the deal were not disclosed.is WME getting a stake in Loti? I’m not sure but I asked WME and am waiting to hear back. Artificial intelligence has been seen as both friend and foe in Hollywood a tool that could potentially make processes more efficient and inspire new innovations, but is also seen as a job killer and yet another way for intellectual property to be stolen. The need for better protections against AI played a central role in last summer’s strikes by the Writers Guild of America and actors guild SAG-AFTRA. On Tuesday, the non -profit Artist Rights Alliance posted an open letter to technology companies demanding that they “stop devaluing” their work, with signatures from 200 musicians
including Billie Eilish and Elvis Costello
. As deep fakes multiply online, agencies are hoping to use AI to stop the bad actors online.
“The worse game of whack-a-mole you are going to play is dealing with the deepfake problem without a technology partner to help you,” said Chris Jacquemin, WME partner and head of digital strategy.
Loti co-founder Luke Arrigoni launched the startup about a year and a half ago. He previous ly ran an artificial intelligence firm called Arricor AI and before that was a data scientist at Creative Artists Agency, WME’s main rival. Arrigoni said Loti
began working with WME about four or five months ago. WME clients give Loti a few photos of themselves from different angles.
They also
record
ed
short audio clips that are then used to help identify unauthorized content
on the internet
. Loti’s software
will
searches the web and reports back to the clients about these unauthorized images and sends take-down requests to the platforms.
“There’s this kind of growing feeling that this is an impossible problem,” Arrigoni said. “There’s this almost adage now where people say, ‘Once it’s on the internet, it’s on the internet forever.’ Our whole company dispels that myth
… We can issue take downs that are actually 95% effective within one day of removing materials
.”
Arrigoni declined to say the financial terms of the partnership or how many WME clients are using Loti’s technology.
Prior to using Loti’s technology, Jacquemin said, his agency would
have to fight the problem of deepfakes on a much more ad-hoc basis. They’ve They’d have to
ask web platforms,
like YouTube and Facebook right?,
to take down unauthorized materials based on what they saw
while browsing
or
what they
heard through their clients, whose fans would flag
questionable doctored
material.
Jacquemin said that in as early as 2022 there were billions of online ads that violated deception policies.
Loti’s technology provides more visibility into the issue. There may be circumstances where not all unauthorized content will be taken down, depending on the client’s wishes.
But at least the performers will know what’s out there.
Back in 2022, platforms like Meta and Google were already dealing with take downs of billions of ads or ad accounts that violated their deception policies, Jacquemin said.
Now, more people in Hollywood are concerned about how newer AI models,
some which in part are trained with publicly available data
, could potentially use copyrighted works.
These
technologies could further blur the lines between what’s real and fake.
If harmful fake content were to be kept up for too long, it could hurt a client’s business opportunities and commercial endorsements.
“They’re so realistic, that it would be hard for most people to know the difference,” Arrigoni said.
This is the latest partnership WME and its parent company Endeavor
have made with anhad partnerships with other
AI-related company. In January, WME partnered with Chicago-based startup Vermillio to
also
protect clients against IP theft, detecting when generative AI content uses a client’s likeness or voice.
WME’s parent company
Endeavor is a minority investor in Speechify, which
makes
text-to-speech technology. Endeavor’s CEO Ari Emanuel used Speechify’s
technology tool
to create a synthetic version of his voice, which gave the opening remarks to an Endeavor earnings call last year.
(On Tuesday, Endeavor
that its largest shareholder, Silver Lake, will take the company private in a deal valuing it at $13 billion.) So far, Loti is self-funded, Arrigoni said. He said he put $1 million into the company himself. The firm is currently in the process of raising an undisclosed amount for a seed round.
