Sixteen U.S. airports equipped with a facial-recognition system

The U.S. Transportation Security Administration (TSA) has begun testing a facial-recognition system for passenger screening at 16 airports. The Washington Post reports.

The pilot project began at Ronald Reagan Washington National Airport (DCA) amid the COVID-19 pandemic in August 2020.

To use the system, a passenger must approach the document-check kiosk and insert their ID into the machine. Then they must look into the camera for five seconds so the program can compare the live image with the passport photo.

The system uses a one-to-one verification method. It compares only the live image from the camera with the photo on the identification without querying a large database.

A human operator oversees the entire screening process and makes decisions in disputed situations.

The TSA noted that the system does not collect photographs, and the verification process is ‘instant’. However, the agency acknowledged that some data may be stored for up to 24 months to allow the technical team to assess the program’s effectiveness.

According to TSA spokesperson Jason Lim, facial recognition has proven more effective than manual processes.

“This technology definitely enhances security. So far, we are very pleased with the machine’s ability to accurately recognize faces,” he said.

The agency added that minor changes in appearance, such as a change of hairstyle, have little negative impact on identity verification.

Rights groups have expressed concern about potential system glitches and possible human-rights violations. According to Albert Fox Kan, founder of the STOP project, minorities have historically faced errors in facial recognition.

“I do not trust the TSA to evaluate the effectiveness of its own facial-recognition systems,” said the rights advocate.

The TSA rejected the concerns and said the technology meets the highest scientific standards.

“Demographic fairness is a serious issue for us, and it is an important element of testing,” Lim said.

In 2023, the TSA plans to roll out the system at all U.S. airports. In the future, the agency aims to develop technology that would scan only the face for identification, without requiring an ID.

Currently, the system is being tested in partnership with Delta Air Lines. The machine compares the passenger’s face to a government database, which typically consists of passport photos.

Kan argued that passengers should have a choice: use facial-recognition kiosks or traditional checks.

In October, there were proposals to deploy drones and AI cameras at Times Square in New York.

In September, the San Francisco City Council voted to grant police access to private surveillance cameras in real time.

Subscribe to ForkLog’s Telegram updates: ForkLog AI — all the news from the AI world!