Congress moves to rein in police facial recognition with court orders and bias testing

What it does

This bill establishes federal rules for how law enforcement can use facial recognition technology. It requires police to get court orders before searching facial recognition databases (with narrow emergency exceptions), prohibits using the technology for mass surveillance or immigration enforcement, mandates accuracy testing to prevent racial bias, and creates reporting and audit requirements so the public can see how often facial recognition is used and whether it works fairly across racial groups.

Why we flagged it

The bill's core function is to regulate law enforcement use of facial recognition technology with court-order requirements, bias testing, and transparency mechanisms. While it affects technology companies, its primary purpose is protecting citizen rights and preventing discriminatory surveillance, not advancing industry interests.

What the text implies

• The 18-month delay before accuracy standards take effect (Section 106) may allow continued use of potentially biased systems during the interim period, creating a window where citizens could be misidentified before safeguards activate.
• The 'derived' definition (Section 3) creates a high bar for suppressing evidence obtained through facial recognition, potentially allowing convictions based on leads that originated from biased or unlawful searches if prosecutors can argue the evidence would have been 'inevitably discovered.'

The full analysis lists 5 implications of this text.

Who stands to gain

facial recognition software vendors (NIST testing creates demand for compliant systems); independent testing firms (operational testing requirement); legal services (civil litigation provisions)