Taylor Swift fake porn. Realistic yet fictionalized images of Gaza. The list of bewildering deepfakes continues, and as deepfake creation tools become easier and cheaper to use, the waves of fakes are hitting faster and harder.
According to a recent Pew Center poll, nearly two-thirds (66%) of Americans say they at least sometimes encounter altered videos or images that are meant to mislead, and 15% answered that they encounter it frequently. In a separate Axios/Syracuse University survey of AI experts, 62% said misinformation is the biggest factor in maintaining news credibility and authenticity in the age of AI-generated content. I answered that it would be a problem.
So what's the answer? do you have?
If you talk to people like Michael Matias, a cybersecurity expert and co-founder and CEO of Clarity, they'll tell you it's a deepfake detector. Matthias founded Clarity in 2022 with Gil Avriel and Natalie Fridman with the goal of developing technology to detect AI-manipulated media (mainly images).
Clarity is one of several large and small vendors competing to develop deepfake detection tools. Others include Reality Defender, which provides a platform that separates text, video, and image deepfakes, and Sentinel, which focuses on deepfake images and videos.
In fact, it's hard to tell Clarity's products apart from others, at least for this writer. Like competing vendors, Clarity maintains scanning tools available through apps and APIs. The tool leverages several AI models to compare uploaded media (video, images, audio) against a database of deepfakes and AI-generated images. Additionally, Clarity offers a form of watermark that customers can use to indicate that content is legitimate.
But Mathias argues that the differentiator lies beneath the surface, not on the surface, and that Clarity is quickly adapting to new types of deepfakes.
“Clarity leverages AI at its core, but operates as a cybersecurity company,” Mathias said. “Clarity treats deepfakes as viruses, acting like pathogens that fork and replicate rapidly. As such, its solutions are also built to fork and replicate to remain adaptable and resilient. The team built infrastructure and AI models specifically designed to accomplish the requirements.”
Of course, the accuracy of deepfake detection areas is a moving target. Even with the best expertise and technology stack money can buy, it's an impossible game to win given the rate of progress of his deepfake creation app, GenAI. This is likely why some major companies, such as Google, Microsoft, and AWS, are adopting more advanced watermarking and provenance metadata as an imperfect, if imperfect, alternative to deepfake protection.
Regardless, Clarity has had no trouble attracting support. The 13-person New York-based startup recently raised a $16 million seed round co-led by Walden Catalyst Ventures and Bessemer Venture Partners, with participation from Secret Chord Ventures, Ascend Ventures, and Flying Fish Partners. Ended.
And it seems to have carved out a niche market. Clarity, which sells subscription and pay-as-you-go plans, initially sought customers in the public sector, including news publishers and the Israeli government. (Matthias claims that Clarity helps authenticate and verify videos originating from the conflict between Israel and Hamas.) But the service has since been linked to identity verification providers and other anonymous “big companies.” It was also expanded to
“This is a fast-paced arms race, just like traditional cybersecurity,” Mathias said. “Companies that want to tackle deepfakes need to act as quickly as the companies that create and spread them.”