Q&A  | 

Clare Garvie, Center on Privacy and Technology at Georgetown Law

"Over half of all US citizens are in a face recognition database used for criminal investigations."

Tags: 'algorithm bias' 'civil rights' 'Clare Garvie' 'data protection' 'Face Recognition' 'face surveillance' 'georgetown law' 'Privacy'

SHARE

Clare Garvie joined the Center as a Law Fellow after graduating from Georgetown Law in 2015, and now serves as a Senior Associate. In 2019 she testified before the House Oversight Committee about police use of face recognition. Her commentary has appeared in The New York Times, The Washington Post, and The Wall Street Journal, and she serves as an expert resource to both Democrats and Republicans in Congress and state legislatures. Her current research focuses on the use of face recognition-derived evidence in criminal cases and the ways activists, public defenders, and policymakers can ensure the technology is under control.

Can you give us an overview of your work?

The Center on Privacy & Technology is a think tank that examines how new technologies may create new privacy risks, particularly in historically marginalized or oversurveilled communities.

I research the use of face recognition technology by police departments in the United States and the risks this poses to our various constitutional rights, including free speech and association, privacy, due process, and equal protection.

Within that, I research the use of face recognition technology by police departments in the United States and the risks this poses to our various constitutional rights, including free speech and association, privacy, due process, and equal protection.

How is face recognition present in our everyday life?

Face recognition technology is increasingly used by both public agencies and commercial entities—from police departments to social media and our cell phones. Most driver’s license databases are face recognition databases. We estimate conservatively that a quarter of all law enforcement agencies have access to a face recognition system—the actual number is probably much higher.

Which are its positive uses?

Face recognition is a tool that can help with identification. There are undoubtedly uses for this type of a tool that enhance convenience and public safety. But without safeguards against risky uses or misuse, the potential harms of face recognition are also significant.

In average, how many of us are in face recognition data bases nowadays?

Our research has shown that over half of all US citizens are in a face recognition database used for criminal investigations.

This is because at least 32 states permit police to run or request investigative searches of their DMV systems. All passport and visa photo databases are face recognition databases now too.

This is because at least 32 states permit police to run or request investigative searches of their DMV systems. All passport and visa photo databases are face recognition databases now too.

You have said that face recognition poses very real concerns to our constitutional rights, and may even violate the fourth amendment of the US Constitution. How is that?

We have an expectation of privacy to our movements across time and space even when they occur in public. This is because those movements can betray sensitive information about who we are—when and where we pray, seek medical attention, drop our children off at school, and more. If face recognition is used to track us based on where our face shows up on camera, it may violate that right to privacy.

Is there a way we can protect ourselves from facial recognition?

The best way to protect ourselves from face recognition is to advocate for restrictions or bans on its use.

It is not our responsibility as individuals to affirmatively protect our constitutional rights; rather it is the government’s obligation not to violate those rights.

It is not our responsibility as individuals to affirmatively protect our constitutional rights; rather it is the government’s obligation not to violate those rights.

You've also said that face recognition works less accurately on women, people with darker skin and young people. How is that?

Studies continue to show that face recognition performs differently depending on the demographics of the person being searched, including race, gender, and age. Depending on the goal of the system,

Those differences may lead to a system disproportionality mis-identifying, or failing to identify, certain people, which raises questions about equal protection under the law.

those differences may lead to a system disproportionality mis-identifying, or failing to identify, certain people, which raises questions about equal protection under the law.

Whereas people are concerned about the threat face recognition poses for civil liberties and privacy rights among others, is there a way this technology can support human rights?

Face recognition is an identification tool, so to the extent that the goal of identification is in line with human rights goals, face recognition may support human rights. For example, face recognition could be used to help with citizenship registration, facilitating aid delivery and voting. But it is not without risks—like with any tool, it’s harms and benefits depend on how it used and the checks we have put in place against its misuse.

Face recognition seems to be equally non regulated in all developed countries. How was implementing this technology possible in such circumstances?

To date, the development and deployment of face recognition technology has far exceeded most countries’ ability to regulate such use, leading to widespread, unregulated adoption.