(San Francisco) American taxpayers will soon no longer need to authenticate themselves on the tax site via facial recognition, an artificial intelligence (AI) technology decried for years by human rights defenders.
Posted at 4:04 p.m.
The federal tax agency, the IRS, announced on Monday that it was giving up using ID.me, a facial recognition authentication service.
“The IRS takes the privacy and security of taxpayer data seriously. We understand there are concerns about (the current system),” agency head Chuck Rettig said in a statement Monday.
“Everyone should be comfortable with how their personal information is saved, and we are actively looking for short-term alternatives that don’t require facial recognition,” he added.
Democratic Senator Ron Wyden, who had protested against this system, welcomed this decision on Twitter.
“Forcing Americans to use facial recognition as a condition for interacting with essential government services is unacceptable,” he wrote in a letter to the IRS on Monday.
He also points out that this type of technology has created problems of discrimination, particularly against non-white people.
This is a long-standing criticism of opponents of these AI methods: algorithms, trained from predominantly white populations, make more errors on black people in particular.
Many NGOs and elected politicians therefore accuse facial recognition of systematizing the human biases already present in society, such as racism. They also point out the risks in terms of personal data protection.
China is using large-scale facial recognition cameras.
In the United States, faced with pressure from associations, large groups such as Amazon, Microsoft, IBM and Google have stopped, at least temporarily, selling their facial recognition software to police forces.
And Facebook decided last November to dispense with facial recognition, which, since 2010, has made it possible to identify people present in photos or videos posted on the social network.
A year ago, a federal judge in Illinois approved a settlement between Facebook and individuals who accused the platform of using data related to facial recognition without their consent. The Californian group then agreed to pay them $650 million in damages.
ID.me did not react immediately to a request from AFP.