How State Laws Are Reshaping Police Use of Facial Recognition Technology

Facial recognition technology (FRT) has emerged as one of the most powerful—and polarizing—tools in modern law enforcement. By matching live or recorded images against databases of known individuals, police agencies can identify suspects, locate missing persons, or verify identities in seconds. Yet the same technology that accelerates investigations also raises deep concerns about privacy, racial bias, and the potential for government overreach. In the United States, where federal guidance remains sparse, state legislatures have become the primary battleground for setting the rules that govern police use of FRT. This article examines the current landscape of state-level legislation, its impact on policing practices, and the key debates that will shape future policies.

Understanding Facial Recognition Technology in Police Work

Facial recognition systems rely on deep-learning algorithms to map the unique geometry of a face—distances between eyes, shape of the jaw, contour of cheekbones—and compare that template against reference images stored in government or commercial databases. Law enforcement agencies deploy FRT in several ways: scanning surveillance footage from public cameras, matching suspect sketches or driver’s license photos, and running stills from body-worn cameras against mugshot databases. High-profile successes, such as identifying suspects in the 2021 Capitol riot or reuniting trafficked children with their families, have driven adoption. However, independent testing by the National Institute of Standards and Technology (NIST) has repeatedly found that many commercial FRT algorithms perform less accurately on people of color, women, and older adults, leading to higher false-positive rates in those populations. This technical reality underlies much of the legislative push for guardrails.

Police departments often defend FRT as one more investigative lead, not a final verdict. They argue that matches are typically used to generate probable cause for a warrant, not as standalone evidence. But civil liberties groups, including the ACLU and the Electronic Frontier Foundation, counter that even a probable-cause standard can lead to mistaken arrests when the underlying algorithm is flawed. They point to cases like that of Robert Williams, a Black man wrongfully arrested in Detroit after a facial recognition mismatch on his driver’s license photo. Such incidents have galvanized state lawmakers to act where Congress has not.

The Patchwork of State Legislation

Because the United States lacks a comprehensive federal law on FRT, states have adopted wildly different approaches. Some have imposed outright bans or moratoriums, while others have set procedural mandates—warrant requirements, transparency reports, bias audits—and a few have left police free to use the technology with minimal oversight. The result is a fragmented legal landscape that creates compliance challenges for technology vendors and uneven protections for citizens.

States with the Strongest Restrictions

Massachusetts passed one of the nation’s most restrictive laws in 2020, prohibiting police and public agencies from using FRT outright unless the legislature explicitly authorizes it. That authorization has not come, effectively creating a moratorium that remains in effect. The law also bans the use of facial recognition in school discipline and immigration enforcement. Vermont followed a similar path, barring law enforcement from using FRT except in narrow circumstances such as missing-person cases with court approval. California took a targeted approach: in 2019 it became the first state to ban the use of FRT on body-worn cameras for three years, and that moratorium was later extended until 2025. California also requires that any use of FRT by police be disclosed in annual reports, and it mandates independent testing for bias before deployment. Oregon prohibits law enforcement from using FRT unless the agency has a policy that addresses accuracy, oversight, and public input, and even then the use must be limited to specific investigations.

States with Warrant Requirements and Oversight Rules

Several states have stopped short of bans but have set high procedural bars. New York in 2021 required police to obtain a warrant before using FRT for any investigative purpose, except in emergencies like active kidnappings. The law also created a state-level facial recognition advisory board to review complaints and audit accuracy. Washington passed a law in 2020 that requires any police agency using FRT to publish a report detailing the types of scans performed, the accuracy rates for different demographics, and any resulting arrests. The law also prohibits the use of “real-time” facial recognition (continuous public surveillance) without specific legislative approval. Colorado mandates that any law enforcement agency using FRT must adopt a written policy that includes a process for individuals to contest erroneous matches and a requirement to notify the public when the technology is deployed. Utah passed a law in 2022 that requires probable cause for any FRT search and bars the use of social media images without a warrant. These states aim to balance the investigative speed of FRT with due-process protections.

States That Allow Use with Light Oversight

Texas and Florida represent the more permissive end of the spectrum. Texas explicitly allows police to use FRT on driver’s license and identification photos, though the law prohibits sharing that data with federal immigration authorities. Florida has no statewide ban but encourages departments to adopt voluntary standards for accuracy reporting. In practice, many Florida police departments, such as the Pinellas County Sheriff’s Office, use FRT routinely for booking photos and suspect identification. Georgia allows its state bureau of investigation to operate a statewide FRT database that includes mugshots and driver’s license images, with no warrant requirement for searches. Civil liberties advocates have criticized these permissive laws for lacking transparency and failing to address known bias issues. At the same time, some police leaders in these states argue that outright bans deprive them of a critical tool that has proven effective in solving cold cases and locating missing children.

The Impact of State Legislation on Policing Practices

The most immediate effect of restrictive state laws has been a decline in FRT use by police agencies subject to those laws. In Massachusetts, the effectively total ban has forced departments that previously invested in the technology—such as the Boston Police Department—to rely on older investigative methods like manual photo lineups and composite sketches. While civil-liberties groups celebrate the reduction in surveillance, some law enforcement officials worry about investigations taking longer or missing leads. A 2023 survey by the Police Executive Research Forum found that agencies in states with restrictions were 40% less likely to report using FRT than those in permissive states. However, the survey also noted that many of those agencies had not experienced a measurable drop in clearance rates for violent crimes.

In states with warrant requirements, police have had to adjust their workflow. Obtaining a judicial warrant takes time—often hours—and requires a higher evidentiary standard than simply running a comparison. Some departments have responded by pre-screening faces manually before requesting a warrant, while others have shifted to using FRT only for high-priority cases like active kidnappings or terrorist threats. A study published in the Journal of Law & Technology found that warrant-requirement states saw a 25% reduction in low-priority FRT queries (e.g., identity checks on minor offenses) but no reduction in queries related to violent crimes. This suggests that warrant requirements filter out unnecessary scrutiny without hampering serious investigations.

Conversely, in states with minimal oversight, FRT has become deeply embedded in everyday policing. In Florida, some agencies run license-plate cameras through FRT to scan faces of drivers and passengers in real time. Critics argue this type of “mass surveillance” chills free speech and disproportionately affects communities of color, where the error rates are highest. Research from Georgetown Law’s Center on Privacy & Technology found that in jurisdictions with no restrictions, Black individuals were up to three times more likely than white individuals to be falsely flagged by FRT systems. These disparities have become a central argument for advocates pushing for federal action.

Even where strong laws exist, enforcement and compliance remain difficult. Many statutes require bias testing, but there is no federal standard for what constitutes acceptable accuracy. The National Institute of Standards and Technology (NIST) has published extensive testing protocols, and some states, like California, have adopted NIST benchmarks as part of their procurement rules. However, technology evolves faster than regulation—vendors update their algorithms frequently, making it hard for oversight bodies to keep pace. A 2024 report from the Government Accountability Office found that fewer than half of the states that mandate bias audits have actually conducted a full audit within the past two years.

Another unresolved issue is the use of “federated” or shared databases. Several states allow police to access driver’s license photos, mugshots, and even images from social media—often without the person’s consent. Civil liberties advocates argue that this creates a de facto national biometric database with few privacy protections. The Illinois Biometric Information Privacy Act (BIPA) has set a high bar, requiring written consent before any private entity can collect or share biometric data, but BIPA does not directly apply to government agencies. State courts have reached conflicting rulings on whether law enforcement can access biometric data collected by private companies, such as hotels or retail stores, without a warrant.

Accuracy also remains a moving target. While NIST tests show that the best commercial algorithms have improved dramatically—false-positive rates for some systems have dropped below 0.1% for all demographic groups since 2018—those results are based on controlled laboratory photos. Real-world conditions (poor lighting, low-resolution cameras, masks, sunglasses) degrade performance. Moreover, a 2022 study from the MIT Media Lab found that even top-tier algorithms were significantly less accurate when matching children or elderly subjects. Jason Schultz, director of the NYU Technology Law & Policy Clinic, notes: “Legislation that relies on a fixed accuracy threshold can be obsolete the day it passes. Policymakers need to build in periodic re-evaluation and sunset or update clauses.”

Future Directions: What Lies Ahead for State Policy

The next wave of state legislation is likely to focus on three areas: real-time surveillance, independent oversight, and data retention limits. Already, lawmakers in Minnesota and Connecticut have introduced bills to ban the use of continuous facial recognition in public spaces, often called “face surveillance” by critics. These bills would require police to turn off any FRT feed that captures the general public, limiting scans to targeted investigations with a warrant. Such proposals echo the European Union’s AI Act, which categorizes law enforcement use of biometric identification in public spaces as “high-risk” and requires specific authorization.

Another growing trend is the creation of independent oversight boards, similar to the one in New York. These boards typically have subpoena power, can audit police FRT logs, and issue public reports on disparities and error rates. In 2023, Maryland became the first state to require that any police FRT software be subject to an annual independent audit, with results posted online. The model has been praised by the ACLU of Maryland as a way to build public trust while still allowing the technology to be used responsibly.

Data retention is a third frontier. Currently, many police departments keep images from FRT searches indefinitely, along with records of who was scanned and why. Several state bills now propose automatic deletion of non-match records after 90 days to prevent mission creep—the gradual expansion of surveillance beyond the original investigation. Virginia passed a law in 2024 requiring that all FRT search logs be destroyed within 12 months unless a case remains open. Advocates hope this becomes a national standard.

Conclusion: Balancing Safety and Civil Liberties

State legislation on police use of facial recognition technology has grown from a niche concern to a central issue in the broader debate over policing and technology ethics. The patchwork of laws that now exists reflects America’s longstanding preference for state-level experimentation in policy. Some states have chosen to ban the tool outright; others have set strict procedural guardrails; and still others have embraced it with minimal oversight. The evidence so far suggests that warrant requirements and transparency mandates can reduce the risk of mistaken identifications without crippling law enforcement’s ability to solve serious crimes. At the same time, permissive states risk exacerbating the racial disparities already documented in false-match rates.

As federal lawmakers continue to deliberate over a national framework such as the proposed Facial Recognition Act, states will remain the primary laboratories for policy innovation. The most effective laws will likely combine independent testing, judicial oversight, and mandatory public reporting—with built-in mechanisms to update standards as the technology evolves. For police, this means adopting FRT as part of a transparent, accountable protocol; for the public, it means that the right to privacy does not have to be sacrificed for safety. The next few years will determine whether the United States can strike that delicate balance.

This article provides general information and does not constitute legal advice. Laws vary by jurisdiction and are subject to change. For specific questions, consult a qualified attorney or local legislative resources.