Amanda Watford, States Newsroom

Flock Safety is tightening privacy and oversight controls for its nationwide network of automated license plate readers as more communities pull back from the technology and lawmakers across the political spectrum push for new restrictions.

The company announced Thursday that it will recommend that police agencies retain license plate reader data for seven days, down from 30, while adding tools to restrict data sharing between agencies and detect potentially abusive searches. Flock also plans to require law enforcement users to associate searches with case codes and automatically suspend users whose activity is flagged as abnormal.

The changes come amid growing scrutiny of a surveillance network that Flock says has more than 120,000 cameras across 49 states and serves thousands of law enforcement agencies and communities.

The cameras photograph passing vehicles and record information including license plates, vehicle characteristics, location and time. Police can search the resulting data during investigations, sometimes across jurisdictional lines.

More than 50 agencies or communities have canceled, suspended or rejected a Flock contract or deactivated their cameras since the beginning of 2026, according to DeFlock, a grassroots group that tracks license plate reader deployments. Cameras have also been vandalized in communities around the country.

Recent incidents involving police officers accused of using Flock to track friends, relatives and romantic partners have added to concerns about how the data can be accessed and shared. The technology also has been criticized after inaccurate license plate information contributed to wrongful police stops, arrests and detentions.

Flock says the new measures will give communities greater control over their data while strengthening safeguards against misuse.

“These changes, like I said, are the next chapter of a long book I hope to write on how we build a better version of America that’s both safer, protects our privacy and holds those who have power accountable,” Flock Safety co-founder and CEO Garrett Langley said in a video statement.

But civil liberties advocates say some of the safeguards have been tried before and question whether the company’s changes address broader concerns about a system that can create a record of people’s movements even when they are not suspected of a crime.

States have taken different approaches to regulating technologies used in policing, including artificial intelligence, facial recognition, drones and automated license plate readers. More than a dozen states have enacted laws governing such technologies, according to the National Conference of State Legislatures, but there is no consistent national framework for their use.

Strengthening privacy and accountability

Flock’s seven-day retention recommendation is one of its most significant changes.

When the company launched nine years ago, it set 30 days as the default recommendation, while allowing communities to choose shorter or longer periods.

For investigations that require older information, Flock is introducing “Evidence Mode,” which will allow detectives to preserve specific license plate data as evidence in an active case.

The change does not automatically apply to existing customers, which can keep their current retention policies. Communities also will continue to choose their own retention periods.

The American Civil Liberties Union called the shorter retention period a potential step forward but said its impact will depend on how the new mode works. In a statement, the group recommended a 48-hour retention period and warned that if searches cause data to be preserved, the new system could allow significant amounts of location information to remain available longer than seven days.

Flock also is giving communities more control over which outside agencies can search their cameras. Under the new “offense filtering” system, a community could allow another agency to search its cameras for specific crimes or offenses and block searches related to immigration enforcement, for example.

Flock also will require all law enforcement customers to adopt its audit assistance tool by the end of the year. The system looks for unusual search activity and flags it for administrator review. Flock plans to automatically suspend users whose activity meets certain criteria for abnormal behavior while their access is reviewed.

The company also will require case codes for law enforcement searches by the end of the year. Officers can bypass the requirement during emergencies, but those searches also will be flagged for review.

Police misuse and erroneous alerts

Flock’s new controls build on safeguards the company has introduced in recent years, including electronic audit logs and requirements that users provide a reason for searches.

The ACLU, in its statement, said some of those earlier safeguards were ineffective. The group pointed to cases in which officers circumvented search-purpose restrictions by entering vague or meaningless explanations before Flock replaced a free-text field with a list of approved search reasons.

The Washington Post reported earlier this month that nearly 50 police officers had been charged or accused of using Flock cameras for unauthorized purposes, many involving attempts to track current or former romantic partners or family members.

In Savannah, Georgia, four police officers and two civilian employees were fired last week after a department investigation found searches involving personal acquaintances and family members, according to a news release. The investigation also found that an officer had allowed an outside law enforcement agency to access Savannah’s Flock system.

The activity was identified using Flock’s audit assistance tool. All six cases were referred to the Georgia Bureau of Investigation for independent criminal investigations.

The technology has also drawn scrutiny when information in the system is inaccurate or outdated.

In Wisconsin, a woman was pulled over at gunpoint earlier this month after a Flock camera incorrectly flagged her vehicle as being involved in a homicide. Brookfield police stopped Amber Newell on Interstate 94 after receiving the alert, but Milwaukee police told officers they no longer needed the vehicle or its occupants. Newell was released.

Milwaukee police later said the incident was caused by an employee’s failure to remove an outdated wanted-vehicle alert from the Flock system, rather than an error by Flock itself. Newell said she had also been stopped at gunpoint earlier that week and was afraid to drive afterward.In its announcement, Flock said its alerts are intended to be investigative leads rather than conclusions and that officers should visually confirm a vehicle’s license plate before taking action. The company said its system produces fewer than nine human-reported errors for every 1 million alerts.

Bipartisan interest

A bipartisan group of Pennsylvania lawmakers announced Monday that they plan to introduce legislation establishing statewide privacy and accountability requirements for automated license plate readers, including Flock cameras.

The proposal would establish rules governing access to the data, documented reasons for searches, audit trails, data retention and transparency. It also would create civil penalties for unauthorized access, use or disclosure and allow people harmed by misuse to bring civil suits.

“Technology that can help keep our communities safe should never become a tool for stalking or abuse,” Democratic state Rep. Jennifer O’Mara said in a statement announcing the proposal.

Republicans in Congress also have introduced or plan to introduce at least two proposals aimed at restricting automated surveillance technology. One would restrict federal agencies from purchasing or using certain automated surveillance technology, while the other would bar state and local police from using federal funds to acquire automated license plate reader systems.

Last month, 15 Democratic lawmakers in Michigan sent a letter to the state Department of Transportation to end an agreement allowing federal agencies, including U.S. Border Patrol and Homeland Security Investigations, to install surveillance technology on state property. The lawmakers argued that automated license plate readers infringe on residents’ constitutional rights.

The West Virginia House Democratic Caucus also released a statement last month opposing Flock cameras and said it is monitoring canceled contracts and lawsuits involving the technology.

Abandoning the technology

Some local governments are going further by ending their relationships with Flock altogether.

DeFlock’s tracker lists dozens of communities that have canceled contracts, deactivated cameras or rejected proposals since 2021.

More than 50 agencies or communities have taken one of those steps since the beginning of 2026.

Several communities in Rhode Island and Nebraska have acted this month, with city leaders voting to cancel contracts or uninstall cameras.

Earlier this month, the Burnet County Sheriff’s Office in Texas announced that it had removed all fixed Flock cameras after repeated vandalism damaged the equipment. The cameras were provided by the Heart of Texas Auto Theft Task Force, a multi-agency, multijurisdictional task force in Central Texas. The sheriff’s office has initiated criminal cases, but there are currently no suspects, according to a news release.

The debate over this technology is likely to continue through the rest of the year and into 2027, when more state legislatures return to session. That could bring additional proposals to regulate artificial intelligence tools used in policing, as well as how license plate reader data is collected, retained, shared and used.

Stateline reporter Amanda Watford can be reached at awatford@stateline.org.

This story was originally produced by Stateline, which is part of States Newsroom, a nonprofit news network which includes Louisiana Illuminator, and is supported by grants and a coalition of donors as a 501c(3) public charity.

For the first time in years, there are early signs that police killings in the United States may be declining — after deaths reached a record high in 2024 and amid intensified scrutiny of law enforcement tactics nationwide.

The findings come as photos and videos of aggressive law enforcement — particularly involving federal immigration agents — have dominated headlines and social media. The new numbers don’t include deaths during immigration enforcement, and federal agents operate under different authorities and standards than state and local police. Nevertheless, some experts say the heightened visibility has sharpened public attention on the use of force.

New data from Campaign Zero, a research group that advocates for the end of police violence, shows a slight drop in police killings in 2025 compared with 2024.

At least 1,314 people were killed by police in 2025 — the first annual decrease since 2019, according to the group’s report. By comparison, at least 1,383 people were killed by law enforcement in 2024, the highest number recorded since the group began tracking the data.

Some policing experts caution that it’s too early to say whether the drop is the beginning of a longer-term decline.

“You want to have a couple of good years, and you want to begin to gather why we think these things are happening,” said Tracie Keesee, co-founder of the Center for Policing Equity and an associate professor of public safety and justice at the University of Virginia School of Continuing and Professional Studies. Keesee has 25 years of law enforcement experience.

“What do we not know?” she said. “What’s the data not telling us? I think that’s also important.”

Experts point to a range of possible explanations for the decrease in police-related deaths, including ongoing staffing shortages that have resulted in fewer officers on patrol, expanded use of de-escalation training and stricter use-of-force policies, and the uneven rollout of changes adopted by police departments in the years following the 2020 police murder of George Floyd in Minneapolis.

Lower crime rates nationwide — including a decline in homicides — is another possible factor, some experts say, as it may have reduced the number of high-risk encounters between police and civilians.

The uncertainty reflects long-standing gaps in national policing data. There is no comprehensive federal government database tracking police use of force, leaving the public to rely on independent efforts such as Campaign Zero’s Mapping Police Violence database, which compiles incidents from public records, media reports and other verifiable sources.

Last year, the Trump administration shut down the National Law Enforcement Accountability Database, a system that tracked misconduct by federal law enforcement officers.

The available data that is maintained by the federal government is collected by the FBI through its Uniform Crime Reporting system, which began tracking use-of-force incidents in 2019. The data relies on voluntary, self-reported submissions from police departments.

Another widely cited effort, The Washington Post’s Fatal Force database, tracked fatal police shootings between 2015 and 2024, but stopped updating the numbers in 2025.

While the Fatal Force database focused solely on police shootings, the Mapping Police Violence database takes a broader approach, including deaths involving other types of force as well as some accidental deaths — differences that can shape overall counts and complicate comparisons.

Researchers say these gaps are not just a data problem but also a barrier to understanding use of force itself. The gaps make it difficult to study when and why force is used and to evaluate which policies — whether legislative or within police departments — are the most effective in reducing it.

“There really is a significant misconception about what use of force looks like, and it’s largely because of the fact that we just don’t know what leads to use-of-force incidents,” said Logan Kennedy, an assistant professor of criminal justice and criminology at East Carolina University. “There’s not data out there.”

Variation across states

State-level data from Campaign Zero shows wide variation not only in how often police kill civilians, but also in the types of encounters that turn fatal.

Some states consistently had far lower rates of police killings than others. Rhode Island was the only state that had no police killings in 2025, according to the report.

New Jersey had the second-lowest rate in the country in 2025, with 0.08 police killings per 100,000 people. That’s a 48% decrease from the state’s average of the previous 12 years, according to the report.

By contrast, New Mexico had the highest rate of police killings per capita, with 1.36 police killings per 100,000 people, according to the report.

The types of incidents that lead to deadly force also vary. In some places, fatal encounters are more likely to stem from reported violent crimes, while in others they more often begin with routine traffic stops or calls related to mental health crises or welfare checks, according to Stateline’s analysis of the data.

Some researchers and policing experts say those differences may reflect a mix of factors, including training standards, department policies and whether states have invested in alternatives to traditional policing — such as crisis response teams that handle mental health calls.

Since 2021, every officer in New Jersey has been required to undergo de-escalation training known as ICAT, or Integrating Communication, Assessment and Tactics.

ICAT training teaches patrol officers how to handle tense situations — especially those involving people in crisis — by slowing encounters down, communicating clearly and using safer alternatives to force. The program was developed by the Police Executive Research Forum, a national nonprofit focused on policing standards, about a decade ago.

“In the last 10 years, we have seen the evolution of police training, especially as it relates to de-escalation,” said Chuck Wexler, the group’s executive director.

ICAT has been implemented in roughly 1,500 law enforcement agencies nationwide, Wexler said. He added that it may have contributed to New Jersey’s significant decline in use-of-force deaths in 2025, though he acknowledged it would not have been the sole factor.

At least 12 cities with populations over 250,000 had zero police killings in 2025, according to the report. Departments in two of those cities, Long Beach, California, and Minneapolis, have received ICAT training, Wexler said. Police in Roanoke, Virginia, and Spokane, Washington, reported no officer-involved shootings in 2025, and were also trained under ICAT.

“If you don’t change your training and your tactics and how you communicate with people, you’re not going to see the change in the areas that you can,” Wexler said.

Some states, including California and Washington, have adopted stricter use-of-force laws in recent years, allowing officers to use deadly force only as a last resort. Others have expanded certain programs aimed at reducing police involvement in nonviolent situations, such as when someone is in the midst of a mental health crisis and might be better helped by a specially trained social worker than a responding law enforcement officer.

The report’s authors found no single policy directly linked to lower rates of police killings.

The variation, some policing experts say, highlights how uneven changes to policing standards and procedures have been implemented since Floyd’s death.

Some states and localities have pursued sweeping changes, while others have taken a more limited approach. Some experts say it can take years for a policy or training change to be implemented, take hold and begin to shift broader trends.

It’s also unclear whether the momentum behind policing policy changes has been sustained across much of the country — and to what extent states and localities have maintained those changes or rolled them back, experts say.

“Years later, we don’t really know. Did those reforms actually go into effect?” said Kennedy, of East Carolina University. “Asking questions about whether or not they’re persisting or eroding –– it makes a significant difference.”

Disparities persist

The impact of police violence also remains deeply uneven — both nationally and within states.

Black Americans continue to be killed by police at disproportionately high rates compared with white Americans, a disparity that holds across nearly every state analyzed, according to the report. Nationwide, Black people are killed at more than twice the rate of white people, the report found, with even wider gaps in some states.

Native Hawaiian and Pacific Islander, American Indian and Alaska Native, and Hispanic people were also more likely than white people to be killed by police in 2025, according to the report.

Even if 2025 does mark the start of a new downward trend in police-involved killings, some experts say national figures can obscure what’s happening on the ground.

The decline does not mean all communities are experiencing the same level of change, according to Keesee, of the Center for Policing Equity.

“The question I always ask (is), ‘Police killings are down for who?’” Keesee said. “When you still have racial disparities, that means it might not be perceived that killings are down, especially if you’re in communities where a lot of these things seem to take place.”

Stateline reporter Amanda Watford can be reached at ahernandez@stateline.org.

New Jersey Monitor is part of States Newsroom, a nonprofit news network supported by grants and a coalition of donors as a 501c(3) public charity. New Jersey Monitor maintains editorial independence. Contact Editor Terrence T. McDonald for questions: info@newjerseymonitor.com.