Looking to prevent such indiscriminate use — and the potential for abuse — Maryland lawmakers this session passed a law, to go into effect Oct. 1, that will severely limit use of genealogy websites to the investigation of serious violent crimes, such as murder and sexual assault, and only as a last resort after all other methods fail, including searches of a less intrusive state database. Police will also have to get permission from a judge, and all records must be deleted when an investigation is complete.
The change in law will protect Marylanders from unnecessary intrusion into and use of their personal data. One of the legislation’s sponsors, Baltimore County Democrat Sen. Charles Sydnor III, had proposed an outright ban on use of genealogical databases in a previous session, but that would have gone too far. Law enforcement needs to use all the tools it can to nab criminals who escape capture, but they must do it in a way that doesn’t trample on citizens’ rights. Protection of privacy isn’t the only issue that raises concern. Care also needs to be taken to prevent the perpetuation of racial disparities that already exist in the criminal justice system. A decades-long mass incarceration problem already means African American men disproportionately fill our prisons, and their information likely disproportionately fills government DNA databases as well. Government oversight and limits can help offer an extra layer of protection.
We also see a need for restrictions of the use of facial recognition software in identifying criminals. Baltimore City Councilman Kristerfer Burnett wants to go even further, having recently introduced legislation, which passed unanimously out of the Baltimore City Council Public Safety Committee, that would temporarily ban such technology at all city departments — except the police department, which is under state control — until an audit can be completed to determine how it’s being used. He has reason to be wary given that the technology often misidentifies African Americans. An analysis by the ACLU found that Amazon’s Rekognition tool incorrectly identified 28 members of Congress. Nearly 40% of the false matches were of people of color. An often-cited MIT study found public facial recognition systems analyzed the faces of white men accurately but not Black women. Amazon, Google and Microsoft recently paused or stopped sales of these products to law enforcement after a push from activists.
But we urge lawmakers to be careful about total, permanent bans of technology that could be beneficial down the road, if used responsibly. Authorities used facial recognition to help identify the uncooperative suspect, Jarrod Warren Ramos, in the mass shooting at the Capital Gazette newspaper. It could be used to better secure government buildings, find missing children and exonerate the innocent. And there are many uses beyond the scope of law enforcement. Rather than ban it, the technology should be regulated by both the federal and local government so companies, government agencies and police departments can’t use it indiscriminately.
Why not set up stipulations on when it can be used — such as only after exhausting other methods, like the state legislation does with genealogy databases — and require a judge’s permission? Require consent for use of the data and only for specified purposes, and don’t allow sharing of the data between departments. Make technology companies prove they have tested their software on a diverse group of people before considering it. Mr. Burnett’s bill would require an annual report of surveillance technology use, which would offer some oversight. That’s the smart way to move forward, with measures for accountability in place. Facial recognition technology needs improvement; there’s no doubt about. But it’s potential as a police tool, once the technology improves and bias issues are fixed, is still unknown. We need more information before ruling it out. We should proceed with caution.
©2021 Baltimore Sun. Distributed by Tribune Content Agency, LLC.
