Facial Recognition Act of 2025
- Sponsor
- Rep. Lieu, Ted [D-CA-36]
- Committees
- Science, Space, and Technology Committee (primary)
- Last action
- Jul 23, 2025
Bottom line
This bill establishes a comprehensive federal framework to restrict and oversee law enforcement's use of facial recognition technology, aiming to balance public safety with civil liberties and address accuracy concerns.
What it actually does
This bill establishes a comprehensive federal framework to regulate law enforcement's use of facial recognition technology. It mandates court orders for most uses, prohibits certain applications like real-time surveillance or use with body cameras, and requires rigorous accuracy and bias testing. The bill also sets forth robust reporting, auditing, and enforcement mechanisms to ensure compliance and protect civil liberties.
Proponents argue
Supporters argue that this bill is essential to protect civil liberties, prevent discriminatory practices, and ensure accountability in the use of powerful facial recognition technology. They contend that without clear legal standards and oversight, this technology poses significant risks to privacy, free speech, and due process, especially given documented issues with algorithmic bias.
Opponents contend
Critics might argue that the bill's stringent warrant requirements, prohibitions on certain uses, and extensive testing mandates could unduly hinder law enforcement's ability to solve crimes and ensure public safety. They may contend that it creates excessive bureaucratic burdens and limits a valuable investigative tool, potentially making communities less safe.
The bill is moderately complex due to its technical definitions and legal procedures, but a focused reader can understand its core provisions within a reasonable timeframe.
SEC. 2
Grant Award Reduction for Non-Compliance
This provision stipulates that if a state or local government fails to substantially comply with the Act's requirements for a fiscal year, the Attorney General must reduce their grant award under a specific federal crime control program by 15 percent in the following fiscal year. This serves as a financial incentive for adherence to the new federal standards.
Supporters argue
Proponents argue this provision is necessary to ensure compliance with the federal regulations, incentivizing states and localities to adopt responsible practices for facial recognition technology.
Critics contend
Opponents might contend that this penalty could disproportionately impact smaller or under-resourced agencies, potentially hindering their ability to provide essential public safety services if funds are cut.
Tradeoffs
This provision balances the federal government's desire for uniform compliance with the potential for financial strain on state and local entities, particularly those with limited budgets.
SEC. 3
Key Definitions for Facial Recognition Use
This section establishes precise definitions for critical terms used throughout the Act, such as 'facial recognition' (automated identification based on facial characteristics), 'face surveillance' (real-time or stored video tracking), 'arrest photo database,' 'reference photo database,' and 'illegitimately obtained information.' These definitions are crucial for determining the scope and applicability of the bill's regulations.
SEC. 101(a) and (b)
Warrant Requirement for Facial Recognition Searches
This provision mandates that law enforcement officers must obtain a court order to use facial recognition in conjunction with a reference photo database, such as those containing driver's licenses or arrest photos. The application for such an order requires probable cause that a serious violent felony has been committed, and an explanation of why other investigative measures have failed. Additionally, arrest photo databases must regularly remove photos of minors, individuals released without charge, or those acquitted.
SEC. 101(c)
Emergency Use and Subsequent Authorization
This provision outlines specific, narrow exceptions to the warrant requirement, allowing immediate facial recognition use in emergencies. These include identifying deceased, incapacitated, or crime victims; AMBER Alert subjects; lawfully arrested individuals during booking; or in situations involving immediate danger of death or serious physical injury. For the latter, subsequent court approval must be sought within 12 hours, and if denied, all obtained information must be destroyed.
SEC. 102(a), (b), (c), (e)
Prohibitions on Discriminatory and Surveillance Uses
This section explicitly forbids several uses of facial recognition: creating records of individuals exercising constitutional rights (like free speech), discriminatory targeting based on protected characteristics (race, gender, etc., with a narrow exception for specific criminal incidents), immigration enforcement, and use with body cameras, dashboard cameras, or drones. It also bans 'face surveillance' (real-time or stored video tracking) and the use of databases containing 'illegitimately obtained information.'
SEC. 102(d)
Prohibition on Sole Reliance for Probable Cause
This provision mandates that a facial recognition match alone cannot establish probable cause for a search, arrest, or any other law enforcement action. Investigative and law enforcement officers are required to examine results carefully, consider the possibility of inaccuracies, and corroborate findings with additional evidence before taking action.
SEC. 104
Comprehensive Reporting on Facial Recognition Use
This section mandates detailed annual reports from state judges, prosecutors, and their federal counterparts on the use of facial recognition. The reports must include information such as the number of orders issued, offenses investigated, and the number of arrests and convictions attributed to facial recognition, disaggregated by race, ethnicity, gender, and age. This aggregated data is then publicly released by the Director of the Administrative Office of the United States Courts.
SEC. 105
Annual Audits and Suspension for Violations
This section requires annual audits of federal, state, and local law enforcement agencies' facial recognition use. Federal agencies are audited by the Government Accountability Office (GAO), while state and local agencies are audited by independent state agencies. The audits aim to prevent misuse and ensure compliance with the Act. If a violation is uncovered, the agency must cease using facial recognition until all violations are corrected, and public notice of the suspension must be given.
SEC. 106
Mandated Accuracy and Bias Testing
This section prohibits law enforcement from using facial recognition systems unless they undergo annual benchmark testing by the National Institute of Standards and Technology (NIST) and operational testing by an independent entity. These tests must assess overall accuracy and identify significant variations based on race, ethnicity, gender, or age. The Assistant Attorney General of the DOJ Civil Rights Division is tasked with issuing a rule to define a 'sufficiently high level of accuracy,' below which systems cannot be used.
SEC. 107
Comprehensive Enforcement Mechanisms
This section establishes robust enforcement provisions for violations of the Act. Evidence obtained in violation is inadmissible in court. Officers who intentionally violate the Act may face administrative discipline. Individuals whose rights are violated can bring civil actions, seeking equitable relief, punitive damages, and statutory damages of $50,000 per violation. The section also allows civil actions for disparate treatment or adverse impact based on protected characteristics.
SEC. 108
Notice to Arrested Individuals
This section requires law enforcement agencies that use facial recognition to identify an arrested individual to provide them with detailed notice. This includes the name of the operating agency and database, a copy of the authorizing order, accuracy/bias reports, probe images used, any modifications made, the candidate list generated by the system, and other relevant police documentation. This information must be provided in an appropriate language if the individual is not fluent in English.
This clause restricts the use of facial recognition warrants to investigations where there is probable cause to believe the person has committed or is committing a 'serious violent felony' as defined in section 3559(c)(2)(F) of title 18, United States Code. This includes offenses like murder, kidnapping, aggravated sexual abuse, and robbery with a dangerous weapon.
SEC. 101(b)(3)(F)
Why it matters:This is standard legislative practice for referencing existing legal definitions rather than reprinting them, ensuring consistency with established federal law.
Case for: Supporters would argue that limiting facial recognition use to serious violent felonies ensures this powerful technology is reserved for the most critical cases, preventing its overreach into minor offenses and protecting civil liberties.
Case against: Critics might contend that this limitation unduly restricts law enforcement's ability to use facial recognition for other significant crimes that may not meet the 'serious violent felony' threshold, potentially hindering investigations.
Estimated impact: Significantly narrows the scope of criminal investigations where facial recognition can be used under a warrant, focusing it on the most severe crimes.
This provision clarifies that the federal warrant authority established in the bill does not automatically grant federal or state law enforcement access to state-maintained reference photo databases (like DMV photos) unless the specific state's law explicitly and unambiguously permits such access and facial recognition searches.
SEC. 101(d)
Why it matters:This is a common legislative technique to clarify the relationship between federal and state law, ensuring states retain certain powers and preventing unintended preemption.
Case for: Proponents would argue this preserves state sovereignty and allows states to set their own, potentially more stringent, privacy standards regarding their citizens' data, reflecting local values and priorities.
Case against: Opponents might argue it creates a patchwork of laws, making federal law enforcement investigations more complex and potentially less effective in states with restrictive laws, hindering national security or interstate crime efforts.
Estimated impact: Could lead to varying levels of facial recognition access for law enforcement across different states, depending on state legislative action, creating a fragmented regulatory environment.
This section clarifies that the Act does not preempt or override any concurrent or more stringent limitations on facial recognition use, or other privacy, civil rights, and civil liberties laws and rules, established by federal, state, or local governments. It also reiterates that the Act does not authorize state or local use of facial recognition unless expressly authorized by state or local law.
SEC. 202
Why it matters:This is a standard way to address federalism concerns and ensure that the federal law sets a floor, not a ceiling, for protections, allowing for stronger local safeguards.
Case for: Supporters would emphasize that this provision allows states and localities to enact stronger privacy protections, fostering innovation in civil liberties safeguards and responding to local community needs.
Case against: Critics might argue that a fragmented regulatory landscape could create confusion for technology developers and law enforcement agencies operating across jurisdictions, potentially hindering consistent application of the law.
Estimated impact: Enables states and localities to implement stricter facial recognition regulations than the federal standard, potentially leading to a diverse regulatory environment across the country.