The developer of the controversial facial-recognition system Clearview AI will be able to obtain a patent for its technology. Politico reports.
The US Patent and Trademark Office has allowed the company to register the system. For this, Clearview AI only needs to pay the administrative fees.
In the patent application the company describes in detail the use of a “web crawler” to obtain images. It notes that online photos linked to a person’s account can help create additional data points for facial-recognition, which can then be used for a machine-learning algorithm.
The CEO Hoan Ton-That said this is the first patent on facial recognition tied to “large-scale internet data.” He added that the service is intended solely for government agencies and the company does not intend to sell the services to other clients.
Clearview AI builds a database for facial-recognition using photographs of people from social networks. In January 2020, the company drew attention thanks to an investigation by The New York Times. Journalists found that the facial-recognition system was used not only by hundreds of law-enforcement agencies in the United States, but also by private firms, including outside the United States.
ForkLog has examined this story and identified the risks to human rights and freedoms posed by facial-recognition systems.
In November, Clearview AI entered the ranks of the most accurate facial-recognition systems.
In October, Hoan Ton-That said that the service’s database had increased to 10 billion processed images.
In July, Clearview AI raised $30 million in a Series B round.
Subscribe to ForkLog’s Telegram channel: ForkLog AI — all the latest news from the world of AI!