NJ lawmakers want safeguards for police use of AI, facial recognition software

Bill sponsors say they want to understand how police use the technology before deciding what restrictions and safeguards are necessary.

Police are turning to facial recognition software to identify suspects and artificial intelligence to write reports, but the use of these technologies by law enforcement agencies raises questions about accuracy, accountability, and the rights of people swept into the criminal justice system.

Two bills in the New Jersey Legislature would require the attorney general’s office to study how police departments across the state use the technologies and recommend statewide policies governing their use.

One bill, A-759, would require the attorney general to produce a report that assesses what kinds of facial recognition technology are available for police use and how accurate they are. The attorney general would be required to create a statewide summary of how law enforcement agencies use the software to generate police reports, then make recommendations for a statewide policy governing its use.

The second bill, A-3076, outlines similar requirements for artificial intelligence but also calls for a statewide policy requiring officers to review AI-generated police reports for accuracy and completeness.

“We need to explore how we can … help public safety departments in New Jersey use this tool and provide early intervention in criminal matters while also maintaining privacy, security, and the public trust,” Joe Danielson, a Somerset County Democrat and chair of the Assembly’s Public Safety and Preparedness Committee, said during an October committee hearing on the bills.

“With growing community awareness around artificial intelligence, proactive communications and awareness of the safeguards become critical to maintain the public trust,” Danielson said.

The stakes are high because police reports, affidavits and other documents often play a central role in criminal prosecutions.

Jason Williams, a justice studies professor at Montclair State University and author of “Policing and the American Racial State from Plantation to Algorithm,” said he’s concerned about police using facial recognition software and AI.

Facial recognition software is not foolproof, he said, and false matches can send officers chasing after an innocent person. Police reports created with or by AI helpers become authoritative accounts that shape legal decisions.

Williams said he’s not opposed to the state studying the technology but added that he’s generally against cops using AI in police work because criminal justice needs to remain completely human-centered.

“AI-generated police reports, to me, are very problematic,” he told The Jersey Vindicator. “A police report is not merely paperwork … and AI-generated language can often omit context, misstate sequences, and make uncertain claims sound settled.”

“Officer review is necessary,” he said. “But it may not be sufficient because sometimes people have this tendency to just rely on the AI output.”

Instead of conducting studies, Williams said he’d rather see the state create meaningful accountability measures, clear restrictions on permissible use, and independent audits.

That includes regularly disclosing how and when police use AI, particularly to defendants who may want to challenge its use or the results. He also wants communities likely to face facial recognition surveillance to have a say in how authorities use the technology.

“If the state is going to move forward with utilizing these tools, I think the community ought to be involved,” Williams said. “Otherwise, this just becomes a new form of governance in the name of so-called ‘efficiency.’”

Other states, such as California and Utah, have already enacted some of the restrictions Williams wants New Jersey to implement.

California, for instance, adopted a law requiring disclaimers on police reports created with the help of artificial intelligence and requiring officers to certify that they checked the reports for accuracy. The law, which took effect Jan. 1, 2026, also requires departments to retain the first AI draft for as long as they keep the final report and says agencies must maintain an audit trail identifying which officers used the technology.

Utah passed a law in 2025 that mandates every agency create an AI policy listing which tools are allowed for what tasks. It also requires disclaimers for AI-generated reports and directs officers to certify that they’ve reviewed the accuracy.

Most other states have not enacted laws governing AI use in police work.

Both New Jersey bills cleared the Assembly Public Safety and Preparedness Committee on Oct. 19.

Democratic Assembly members Al Abdelaziz, Verlina Reynolds-Jackson and Ellen Park sponsored the facial recognition bill. Assembly members Chris Tully, Annette Quijano and Balvir Singh introduced the artificial intelligence measure.

Singh, a Burlington County Democrat, said in an email that many police departments in his district already use AI to help officers write reports, despite the lack of statewide guidance or vetted tools for law enforcement.

“The concern is that police reports are often central to court proceedings, where specific words, phrasing, and details matter,” Singh said. “Generative AI doesn’t just clean up grammar. It can add supporting details, substitute words, and change phrasing for clarity and flow. These unintended changes, especially when they go unnoticed, can alter the meaning of an officer’s account and potentially affect legal proceedings.”

Tully, a Bergen County Democrat, added that AI tools are being aggressively marketed to police as administrative time savers. But lawmakers need to make sure there is always a human in the loop, he said.

“We understand that AI tools are going to be moving very quickly,” Tully said. “Our goal is to make sure we are creating areas of AI governance, especially in areas of public service.”

The attorney general reports that the bills require are the first step in that direction, he said.

“That’s the reason the studies are so important. We want to make sure we’re creating the right laws and regulations to follow,” Tully said. “The state really needs to step up here…we need to create policies that have the proper guardrails to protect the public good.”

State Sen. Benjie Wimberly, a Passaic County Democrat, is expected to introduce identical bills in the Legislature’s upper chamber.

In a Friday email, Wimberly said he proposed the bills because lawmakers have a responsibility to understand how the tools are being used.

“We shouldn’t have to wait for something to go wrong before establishing appropriate safeguards,” Wimberly said. “Public safety and civil liberties are not mutually exclusive. We have an obligation to protect both.”