Just eight months after passing a bill banning local law enforcement agencies from utilizing facial recognition technology, lawmakers have backtracked and lifted the restrictions.
The new bill lifting the ban will allow local and campus law enforcement to purchase and use facial recognition technology for use in certain permissible situations, such as when identifying individuals who police have reasonable suspicion committed a crime.
The law does not apply to Virginia State Police, who use the Centralized Criminal Image system, which allows police to compare images of a suspect with mug shots of criminals.
The legislation would also permit such technology to be used for identifying crime victims or witnesses, victims of sex trafficking and unidentified bodies in morgues.
The bill keeps certain restrictions in place, especially regarding the use of facial recognition technology for surveillance and monitoring, and limits its use in search warrant applications.
Law enforcement was previously criticized for misusing the technology during the 2021 racial justice protests after making arrests based on analysis from street cameras.
However, lawmakers viewed the earlier ban as temporary while lawmakers evaluated the technology more closely.
Democratic Senator Scott Surovell, author of the new bill, said the moderated use of such technology could be an investigatory advantage for police.
“I think it will help police not only solve, but prosecute crime more efficiently,” the senator told ABC News.
Other lawmakers opposed the lifting of the ban, citing the dangers to privacy and civil rights violations. Studies have shown that the software has higher error rates when identifying people of color.
“This technology can be very important to law enforcement for different types of investigatory situations, but it can also be used for a tremendous amount of bad things,” Republican Senator Ryan McDougle said.
“It is not right; it is not as restrictive as it should be,” he added.
Other senators cited the case of Robert Williams, a black man who was arrested by Detroit police after facial recognition technology matched his face with a shoplifting suspect. Williams’ driver’s license photo — kept in a state image database — was erroneously flagged as a match with the suspect, whose image was scanned from blurry surveillance camera footage.
However, supporters of the bill said the new legislation contains “guardrails” for how the technology can be used, including prohibitions on using information obtained from facial recognition when applying for a search or arrest warrant. Police can still use the technology to generate leads, but they must offer corroborating hard evidence to obtain a warrant.
The bill also requires that the technology used by police must be evaluated by the National Institute of Standards and Technology and have an accuracy rating of at least 98% across all demographic groups.
Republican Delegate Glenn Davis said the law does not change much, as police already use publicly available photos on social media sites during investigations.
“All we’re doing here is instead of having law enforcement officers sit there and look through hundreds of photos to try to make that match, we use this technology,” he said.