EXCLUSIVE Facial recognition company Clearview AI is seeking its first big deals, research chief reveals

An illustrative image shows text projected onto a woman’s face in Berlin, June 12, 2013. REUTERS/Pawel Kopczynski/File Photo

Join now for FREE unlimited access to Reuters.com

Register

NEW YORK, Feb 22 (Reuters) – Clearview AI, whose face search engine has become an unparalleled law enforcement tool, is aiming to win its first big U.S. government contracts this year and grow its team by a third as even as the startup faces challenges in courts and Congress, its chief executive told Reuters.

In an exclusive interview, CEO Hoan Ton-That said he would focus on securing deals with federal agencies. Many of the five-year-old New York company’s 3,100 clients are on trial, he said.

“We know some of these agencies are very successful, but they’re just a small five-figure buy or a six-figure buy away,” Ton-That said, referring to the annual deal value. . “And so it’s, ‘Can we get some of those for seven-figure buys, maybe eight-figure? “”

Join now for FREE unlimited access to Reuters.com

Register

Ton-That also revealed a member of Clearview’s technology team for the first time. While rivals have promoted their benches of scientists or published breakthrough articles, Clearview has remained silent – calling itself “Made in the USA” but offering no specifics.

Ton-That said Terence Liu is the Pennsylvania-based computational physicist behind some of Clearview’s algorithms and its vice president of research. They are listed together on a patent application filed Tuesday.

Overall, the nearly 50-person, mostly remote company aims to hire 18 more this year, including adding five engineers to a team of 12, Ton-That said.

Among its plans this year, Clearview will add enhancement tools to clean up research photos and potentially AI to generate younger and older representations so that one day older people can be matched with childhood photos. .

Clearview has become one of the most prominent facial recognition developers over the past two years as it sends matches to authorities from an ever-growing database of more than 10 billion photos that it found publicly published on the Internet. The results link to websites such as Flickr, Instagram, and Getty Images, where captions or other details reveal more about the topics.

Matches of facial recognition tools used by law enforcement in China, India and many other countries typically come from more limited government arrest or identification databases. Police say Clearview’s web search tool helps them find people outside of those datasets.

Although Clearview compares itself to Google Images search, critics say it violates privacy standards and portends more egregious surveillance. Some lawmakers want to ban it.

Data protection authorities in at least four countries, including Canada and France, have said the photo collection breaches privacy laws, and Clearview is fighting lawsuits in the United States that might force him to change tactics. Read more

Giant companies such as Alphabet’s Google (GOOGL.O) and Meta Platform’s Facebook (FB.O) with the data needed to develop competing tools have pulled out of facial recognition, citing societal concerns and the need for clarity regulators. No well-known startup has ventured into the same gray area as Clearview, which raised around $37 million from investors and now wants more.

The Federal Bureau of Investigation, Immigration and Customs Enforcement and the Fish and Wildlife Service are among a dozen US agencies that have used Clearview, a government audit discovered last August. Clearview also won about $50,000 to research augmented reality glasses with facial recognition to secure Air Force base checkpoints.

Critics are alarmed as Clearview plans to enter new industries.

“Clearview AI has a pattern of deception: the company has publicly defended its mass surveillance by claiming it will only sell to law enforcement while privately offering to expand into finance, retail and entertainment,” said Jack Poulson, executive director of the technology accountability group. Technical survey.

COST REDUCTION

Ton-That started developing facial recognition around 2015 after reading articles such as “DeepFace” and “FaceNet” published by scientists from Google, Facebook and top universities showing dramatic advancements in the technology. He reproduced the results, collected photos online, and improved the accuracy to 99% from 70%.

Tuesday’s patent filing covers Clearview’s process for fast, low-cost facial recognition training. Clearview learns from 70 million online photos of all types of people, compared to smaller celebrity photo databases that feed rival systems, Ton-That said.

Clearview uses AI to apply masks, glasses, and other distortions to training images, allowing it to recognize faces when they’re masked, in profile, in the background, or 20 years younger. In turn, Clearview had greater accuracy than competing tools, according to a US government assessment last year.

Research manager Liu officially joined last year after working as a senior software engineer at Bloomberg LP since 2017, he said. He previously partnered with Ton-That to advance the Clearview prototype.

“It was just something hard to convey again,” Liu said.

Other technical staff include a former Coinbase (COIN.O) employee who handles cybersecurity and an entrepreneur who sold his search engine to Clearview in 2019 and oversees data collection. Ton-That declined to name them.

The team’s efforts to rework Ton-That’s original code, including creating two custom database technologies, has reduced per-side processing costs in Clearview’s system by 95% since 2018, Ton-That said. That.

As that closely tracked “cost per face” drops, Clearview can gather additional photos and become more likely to return a lead, he said.

Join now for FREE unlimited access to Reuters.com

Register

Reporting by Paresh Dave in New York and Jeffrey Dastin in Palo Alto, California; Editing by Kenneth Li and Lisa Shumaker

Our standards: The Thomson Reuters Trust Principles.

Comments are closed.