Jeffrey Dastin and Paresh David

NEW YORK (Reuters). Clearview AI has announced that it is aiming this year to win the first large U.S. government contracts. It also plans on expanding its team by a further third while the startup faces legal and legislative challenges.

Hoan Ton That, CEO said in an exclusive interview that his focus will be on close deals with federal agencies. He said that many of the 3100 clients the New York-based company boasts about are only trial customers.

Ton-That refers to the annual deal value. So it is, “Can you get a few more of those to seven-figures or maybe eight-figure buys?”

Clearview technology manager Ton-That was also revealed to the world for first time. Clearview, despite being praised by rivals for their scientific benches and publishing papers about breakthroughs in science, stayed silent – calling itself “Made in the USA” without providing details.

Ton-That indicated that Terence Liu (a Pennsylvania-based computational biologist) is responsible for Clearview’s algorithmic designs and vice president of research. On Tuesday, a patent application was filed that lists them together.

The company, which employs nearly 50 people, is aiming to add 18 to its workforce this year. Ton-That added that they also plan to add five engineers to their team of 12.

Clearview’s plans for this year include enhancement tools that clean search photos, and possibly AI to create younger and older depictions. This will allow seniors to be matched with childhood photos.

Clearview, which has been one of Silicon Valley’s most well-known developers of facial detection over the past 2 years, sends matching authorities from its ever-growing collection of 10 billion photos. This database includes all images that have been posted online. The results link to Getty Images and Flickr. There, captions and other information reveal more details about the subjects.

The majority of matches from facial recognition software used by law enforcement in China, India and other countries come from smaller arrest records or government IDs. Clearview’s search engine helps police locate people who are not in those databanks, according to the Police.

Clearview is compared to Google Images, but critics say Clearview violates privacy rules and will lead to more severe surveillance. It is being opposed by some lawmakers.

Clearview faces lawsuits from the United States to force them to alter their tactics. Clearview has been accused of violating privacy laws in at least four other countries, including Canada and France.

Alphabet’s Google (NASDAQ) and Meta Platform’s Facebook(NASDAQ:), giant companies with data for competing tools, have pulled out of facial recognition. They cited societal concerns as well as the need to be clear from regulators. Clearview raised $37 million and is seeking more. Clearview was the first startup to venture into this gray area.

Clearview is a federal audit which was found in August. It has been used by the Federal Bureau of Investigation and Immigration and Customs Enforcement as well as Fish and Wildlife Service. Clearview won $50,000 to study augmented reality glasses using facial recognition in order to protect Air Force Base Checkpoints.

Clearview considers new industry entry, raising alarm among critics

“Clearview AI has a pattern of deception: the company has been publicly defending its mass surveillance by claiming it will only sell to law enforcement while privately pitching an expansion into finance, retail and entertainment,” said Jack Poulson, executive director of tech accountability group Tech Inquiry.

RECOVERY OF COSTS

Ton-That developed facial recognition in 2015, after studying papers like “DeepFace”, “FaceNet”, and other research from top universities and scientists. He reproduced the results and collected online photos to increase accuracy from 70% to 99%.

Clearview filed Tuesday’s patent application. It describes Clearview’s fast, low-cost process of training facial recognition. Clearview can learn from more than 70 million photos online of different types of people, as opposed to smaller celebrity photo databases that are used in rival systems. Ton-That stated.

Clearview employs AI to create masks and glasses to train images. This allows Clearview to recognise faces that are obscured or deep in the background, 20 years older, or in profile. According to an evaluation by the U.S. government last year, Clearview was able to recognize faces with greater precision than other tools.

Liu, the research chief, officially joined Bloomberg LP last year. He had been a senior software engineering at Bloomberg LP since 2017. Clearview’s prototype was previously developed by him and Ton-That.

Liu stated that it was difficult for her to go on with the same thing again.

An ex-Coinbase (NASDAQ;) employee now runs cybersecurity. Another entrepreneur, who in 2019 sold Clearview his search engine and is responsible for data collection and management, also works as a technical staff. Ton-That did not name them.

Ton-That noted that Clearview has seen 95% lower processing costs since 2018 due to the efforts of the Clearview team in rework Ton’That’s code.

Clearview will begin to collect additional photos as the “cost per facial” drops, and Clearview is more likely than ever to turn a lead.