There’s no better way to decrease public trust in law enforcement, especially during a time when police actions nationwide are being intensely scrutinized following the death of George Floyd, than by lying to the public, as the Virginia State Police did.
The backlash came swiftly last month when the General Assembly overwhelmingly passed a bipartisan bill patroned by Del. Lashrecse Aird, D-Petersburg, that bans the use of facial recognition software by law enforcement and campus police agencies in Virginia without prior express approval by the legislature—making it one of the most restrictive bans on AI in the nation.
The legislature also passed a subsequent amendment by Gov. Ralph Northam that excludes commercial airport police and ironically enough, the Virginia State Police. If the governor signs HB 2031, as he is likely to do, the new law will go into effect on July 1 of this year.
At some point, the law will have to evolve to allow the use of some facial recognition software by law enforcement to help identify victims and solve crimes, but not without a robust public debate. And the technology itself also needs to evolve. The National Institute for Standards and Technology evaluated a number of facial recognition systems in 2019 and found high false-positives for women, especially Black women in particular, and African Americans in general.
Lawmakers will first have to devise strict rules for its use in very limited circumstances. Forbidding most law enforcement agencies in Virginia from using facial recognition software without permission from the state legislature is a good start, although it no doubt won’t be the end of the discussion.
Credit: Source link