Clearview AI wants its facial-recognition tech in banks, schools, etc
I get knocked down but I get up again, Italy, Canada, UK, ACLU, Facebook, Google, YouTube, Twitter... are never gonna keep me down
Clearview AI is reportedly expanding its facial-recognition services beyond law enforcement to include private industries, such as banking and education, amid mounting pressure from regulators, Big Tech, and privacy campaigners.
The New York-based startup's gigantic database contains more than 20 billion photos scraped from public social media accounts and websites. The database was used to train Clearview's software, which works by performing a face-matching algorithm between input images and ones stored on its database to identify individuals.
These images were downloaded without explicit permission from netizens or companies. Although Clearview has been sent numerous cease and desist letters from Twitter, YouTube, Google, Facebook and more, it continued to collect more images and grow its database. The demands to stop scraping public-facing webpages, however, were not legally binding, unlike the settlement agreement Clearview entered into to end its lawsuit against the American Civil Liberties Union.
Clearview promised to stop giving or selling access to its database system to most private companies and organizations across the US. Public agencies and law enforcement, however, can still use its large database. Private sector businesses, instead, can only use data they provide to the company's facial-recognition software; ie, they have to provide their own database of photos. Clearview is also not allowed to use that data to add to its database.
"Clearview AI doesn't use any private images from its customers or anywhere else to train its bias-free facial recognition algorithm," its CEO Hoan Ton-That confirmed to The Register. "Clearview AI only uses public images from the open internet to train its bias-free facial recognition algorithm."
Ton-That had claimed his company's software was only being used by law enforcement to help identify suspects in criminal cases. But Clearview has ambitions to expand beyond those capabilities, and is hoping to provide facial-recognition technology to banking apps and schools, according to Reuters.
"Clearview AI is interested in using facial recognition as a way to help prevent crime and financial fraud. Today facial recognition is already being used to unlock your phone, provide access to buildings, identity checks and even for payments," Ton-That told us.
"Clearview AI provides its facial recognition technology, without the large database of 20B+ images, through Clearview Consent to a visitor management software provider, who provides visitor management services to customers, some of which are schools," he added.
- Clearview AI fined millions in the UK: No 'lawful reason' to collect Brits' images
- Clearview AI promises not to sell face-recognition database to most US businesses
- Ukraine uses Clearview AI to identify slain Russian soldiers
- Machine-learning models more powerful, toxic than ever
The UK's Information Commissioner's Office issued a £7.5 million ($9.43 million) fine for violating the country's data privacy laws, and ordered Clearview to stop scraping photos and delete existing images of its residents, this week.
Still, the company believes its technology is beneficial despite risks of misidentification or issues of data privacy and security. "Facial recognition can be used to help prevent identity theft and fraud. For example, before a large bank transaction, it may be useful to make a facial recognition check with the owner of the account, to ensure that money is not stolen," Ton-That said.
"The potential of facial recognition technology to make our communities safer and commerce secure is just beginning to be realized," he added. ®