Over the past year, the debate over facial recognition technology has heated up. On one hand, facial recognition technology, like other forms of biometric identification, can greatly benefit systems in which the user’s identity must be flawlessly verified. Ravi Raj from Passage AI explains, “Increasingly we’ve seen leaks of sensitive private information including credit card numbers, passwords and social security numbers, through data hacks, often resulting in identity theft. Facial recognition technology can increase the security of sensitive accounts by requiring a biometric scan to access an account in place of a password.”
Facial Recognition Technology Use Cases
Raj says when it’s used ethically and accurately, facial recognition technology can also provide consumers with greater convenience. Raj says, for example, “Facial recognition could be used to identify a passenger when they use public transit and automatically debit their accounts.”
And there are numerous other uses or potential applications for facial recognition technology.
Identifying people in photos posted on social media is a familiar use case. This technology can also provide an efficient and secure solution to unlock mobile devices, target dynamic advertising based on a consumer’s age and gender, automatically track school attendance, and streamline airline check-in processes.
Risks Associated With Facial Recognition Technology
While some industries are exploring the promising potential, there is also ample concern over how facial recognition technology could create risks — so much so that some legislators have passed laws to ban or limit its use, including San Francisco, Oakland, and the Boston suburb of Somerville.
Raj says concerns over facial recognition technology center on three general areas:
Raj says if businesses, enterprises or organizations use facial recognition for applications beyond what they publicly disclosed, it could lead to privacy and ethical issues.
He points out, however, “Consumers do not have to give up privacy as long as their data is used solely for the reason described and nothing else.”
- Data protection
“Beyond privacy,” Raj says, “security of the data is very critical to ensure that hackers and other bad actors don’t make malicious use of the data.”
A report on the 2019 BioStar2 security breach brought some of these issues front and center. One of the biggest concerns is that unsecured biometric data can’t be changed once it’s stolen — it’s easier to update a password than change a face. Also, if account and personal information are stolen with biometric data, cybercriminals can take over accounts — even exchange the account owner’s facial and fingerprint records for their own.
- Racial bias
Raj comments, “There could also be bias in the data that is used for training the deep learning models for facial recognition. Since the data sets used in a particular country would underrepresent minorities, it could lead to bias and algorithm errors when dealing with recognition of minorities.”
Potential for Misuse
The American Civil Liberties Union (ACLU) also raises questions about the use of facial recognition technology for surveillance, which could be carried out without people’s knowledge or consent. The ACLU warns that driver license photographs or other images could be used with surveillance systems to build systems that can track people.
The American Bar Association points out that using facial recognition technology for some use cases could violate the Fourth Amendment, which protects US citizens from unlawful search in places where people have a reasonable expectation of privacy. In 2018, the Supreme Court ruled that collecting historical cell site location information (CSLI) from cellular providers, which could be used to track a person, required a warrant. In its decision, the Supreme Court stated that as technology advances, courts would have to work to protect people’s privacy.
Facial recognition technology could also interfere with First Amendment rights to freedom of association — if people know they’re being watched, they may begin to self-censor their activity.
What Developers Need to Know About Facial Recognition Technology
If you are considering providing your users with identity, user authentication, or access control solutions using facial recognition technology, Raj advises you to vet solutions carefully.
“It is extremely important that facial recognition technology be 100 percent accurate,” he says. “Even a small amount of inaccuracy can lead to inconvenience for consumers, and worse, violation of civil rights and the likelihood of innocent individuals being punished, especially in law enforcement applications.”