In early January of this year, the Internal Revenue Service (IRS), the U.S. federal government agency that collects income and miscellaneous taxes, implemented recognition through ID.me for U.S. taxpayers who wish to view child tax credit and advance credit payments, expanding this technology to all of its departments and all taxpayers very quickly. In response to the outcry over the use of a third-party service, the IRS announced that it would no longer use ID.me services.

If there is one AI technology that is causing debate and concern, it is facial recognition. While China uses it to repress journalists, foreign students, the Uyghur community and “people of concern”, Europe is trying to regulate its use in public spaces via the RGPD. Yet, many of us use this technology on a daily basis, whether it is to unlock our smartphone or computer, to enter our company, to access our devices and computer applications. Virtually every police department in the world uses it, and more and more cities are installing smart cameras. The issues raised by facial recognition are data use and algorithmic bias.

Bias in facial recognition

Studies have shown that most facial recognition algorithms have “demographic differentials” that can worsen their accuracy depending on a person’s age, gender or race. For example, software trained on predominantly light-skinned subjects is logically less effective when the skin is dark, and this would also be the case if the subject is a woman. Many NGOs and politicians have accused facial recognition of systematizing human prejudices already present in society, such as racism, while pointing out the risks related to the protection of personal data. Under pressure, in the US, Amazon, Microsoft, IBM and Google stopped selling their facial recognition software to police forces and Facebook abandoned facial recognition last November.

The IRS and ID.me facial recognition

The IRS implemented ID.me’s facial recognition solution in early January and was expected to expand it to include viewing all personal tax documents on the IRS website within a few months in an effort to prevent fraud. ID.me is widely used in the U.S., including by other government institutions and is recognized as reliable. To verify one’s identity with ID.me, one simply provides a photo of an identity document (driver’s license, ID card or passport) and takes a selfie with a smartphone or computer via the webcam. As a federally certified identification service provider, ID.me is required to store people’s data. That’s what has alerted privacy advocates and some elected officials. Mike Crapo, a senator, said:

“It is highly likely that with personal information on 70 million individuals, including biometric data, ID.me could be a prime target for cybercriminals, rogue employees and espionage.”

In the face of widespread protests, the IRS is backing away from ID.me’s facial recognition and states:

“The IRS takes taxpayer privacy and security seriously, and we understand the concerns that have been raised .Everyone should feel comfortable with how their personal information is secured, and we are quickly pursuing short-term options that do not involve facial recognition.”

However, the IRS points out that this abandonment could cause “the greatest disruption to taxpayers during filing season” and so the ID.me solution will be abandoned after that. According to its statements, the agency will “quickly develop and bring online an additional authentication process that does not involve facial recognition.”

Translated from USA : L’IRS abandonne finalement la reconnaissance faciale d’ID.me