The Current State of Facial Recognition

August 20, 2020

The use of facial recognition in public safety and public services has long been controversial. The technology offers several benefits but demographic biases have given some communities and developers of the technology pause.

The Pros of Facial Recognition

Law enforcement uses facial recognition to catch criminals or find missing persons. In China, facial recognition located a suspect in a crowd of 60,000 people and found a missing person who was abducted 30 years ago. Airports, retail stores, banks and other locations subject to crime use facial recognition to identify known suspects entering the property.

Amid a pandemic, facial recognition combined with fever detection locates travelers that pose a health risk to the public. Travel history could also indicate individuals in a crowd who recently visited an at-risk area but haven’t been evaluated by health physicians. For the confirmed sick in quarantine, facial recognition ensures individuals don’t leave their homes.

Beyond public safety, facial recognition also improves convenience. Facial recognition check-in has been used by Delta airlines to fast track security checkpoints, replacing the need for passengers to repeatedly provide identification. The service is optional, but data reports up to 98% of Delta passengers choose to use the technology. Facial recognition could also be used in retail stores to pay for purchases.

Facial Recognition Bias Concerns

A primary concern is demographic bias. A recent study revealed Asian and African Americans were up to 100 times more likely to be misidentified by facial recognition than Caucasians. Native Americans had the highest misidentification rate of all ethnicities. Women were more likely to be falsely identified than men, and African American women were most often falsely identified in police investigative searches. When it comes to age, elderly and children are also likely to be inaccurately identified.

In 2018 Microsoft called on the United States Congress to regulate the use of the technology due to privacy concerns. This year, due to racial bias concerns, IBM announced it would no longer develop or sell facial recognition technology until Congress can pass use regulations for law enforcement. Amazon postponed police use of its facial recognition technology for one year. In June, Boston was the second largest city in the world to ban facial recognition (San Francisco banned the technology in 2019). Supporters of the ban citied concerns over racial bias and the infringement of civil liberties.

The Future of Facial Recognition

Facial recognition might overcome its biases if it could train itself using advanced artificial intelligence through neural networks or neuromorphic computing. However, overcoming biases doesn’t alleviate privacy concerns or abuse of technology concerns.

Technological advancements, public opinion and government regulations will largely influence the use of facial recognition in the future.