Microsoft just quietly deleted a facial recognition database of more than 10 million images of around 100,000 people — most of them known celebrities — Engadget reports.
The news comes after Microsoft has actively tried to distance itself from the technology.
“The world is on the threshold of technology that would give a government the ability to follow anyone anywhere,” Brad Smith, the President of Microsoft, warned in November 2018, calling for facial recognition software to be regulated.
The company’s MS Celeb training dataset was lauded as the “largest publicly available one in the world” when it was created in 2016. It was designed to train tools for image captioning and news video analysis, according to Microsoft Research’s paper on the matter.
The images were pulled from Creative Commons databases, but the subjects in the 10 million images were not asked for consent, as the Financial Times reports.
“The site was intended for academic purposes,” read an official statement received by the Financial Times. “It was run by an employee that is no longer with Microsoft and has since been removed.”
Caught in the Act?
The dataset, along with two other massive and very similar databases hosted by Duke and Stanford University researchers, was discovered by Adam Harvey, a Berlin-based artist and researcher.
“Microsoft has exploited the term ‘celebrity’ to include people who merely work online and have a digital identity,” said Harvey in a statement. “Many people in the target list are even vocal critics of the very technology Microsoft is using their name and biometric information to build.”
READ MORE: Microsoft quietly deletes largest public face recognition data set [Financial Times]
More on facial recognition tech: The US Army’s Next Rifle May Use Facial Recognition