Several new startups in Seattle are taking on one of the most pressing issues in tech: the growing problem of deepfakes and likeness issues driven by the rapid adoption of generative AI technology.
The generative AI boom has led to an increase in deepfakes, or digitally manipulated photos, videos, and audio that use someone’s likeness to generate synthetic media.
Deepfakes have already become a problem in schools, elections, entertainment, and various other sectors of society.
“There’s not a lot of control over this,” said Luke Arrigoni, CEO of Loti, a Seattle startup that helps public figures take down deepfakes and fake accounts.
Loti is among a crop of up-and-coming tech companies and nonprofits developing software to help identify deepfakes and protect their likeness.
But while venture capitalists are starting to put cash behind these nascent companies, The Washington Post recently reported that some question their accuracy, and whether they provide a false sense of certainty.
Another potential concern is that the line between what is real and synthetic is “increasingly blurring,” the Reuters Institute noted in a report detailing the limitations of such tools.
After running an AI consulting firm for a decade and working as a data scientist for Creative Artists Agency, Arrigoni started Loti two years ago. The company analyzes photos and audio from users, and then scours the internet for potential deepfakes, using advanced facial and voice recognition technology. The company also sends takedown requests to platforms after identifying illegitimate content.
WME, the talent agency giant based in Los Angeles, recently signed a deal with Loti and is using the platform for its clients.
Arrigoni founded Loti with Rebekah Arrigoni and serial entrepreneur Hirak Chhatbar.
Loti employs about 25 people, split between Seattle, Los Angeles, and outside the U.S.
Certifi AI is also focused on spotting and eliminating deepfakes, but with a particular focus on girls and women who are targeted by bad actors.
Some studies show how women are overwhelmingly targeted in deepfake pornography.
Melissa Hutchins co-founded the company after her own personal experience as a victim of cyberstalking.
Certifi AI is initially focused on serving law enforcement and criminal prosecutions, and helping media platforms quickly recognize and remove inappropriate and illegal deepfake content.
Hutchins previously worked in product management at The Disney Company and Expedia Group.
Previously: After personal experience with cyberstalking, tech founder launches startup battling deepfake content
TrueMedia is developing an AI-powered tool to detect deepfake videos, photos, and audio, aiming to combat political disinformation in the lead-up to the 2024 elections.
The non-partisan, nonprofit organization launched in January and is led by Oren Etzioni, a University of Washington professor and longtime AI specialist.
“Help get the word out so that this incredibly critical election is not swayed by deepfakes,” Etzioni said on stage at the GeekWire Awards last month in Seattle.
The organization is funded by Uber co-founder Garrett Camp through his Camp.org nonprofit foundation.
Previously: As elections loom, TrueMedia.org releases AI deepfake detector to identify fake content
Official AI is building a marketplace designed to help people control their digital likeness and connect them with marketers interested in using AI-generated talent.
“We’re really trying to build the ecosystem for authenticity,” said CEO Dave Siegfried.
The new startup is complementary to firms directly tackling deepfakes, according to Siegfried.
“We believe both approaches are needed for generative AI media to flourish equitably,” he said.
Official AI, which generates revenue in part through licensing fees, was originally incubated inside startup studios run by Pioneer Square Labs and Mudita Venture Partners.
Siegfried previously co-founded a lending startup called LOANtuitive and SkyMode, a mobile e-commerce platform. He co-founded Official AI with longtime engineering leader John Pletka.
