In the United Kingdom, police departments are testing out an artificial intelligence (AI) tool to analyze evidence in cold cases—including the unsolved stabbing murder of a 30-year-old man in 1994—in the hopes of catching any missed clues. The software can sift through video footage, financial transactions, social media, emails and other documents simultaneously. Authorities claim the tool was able to review evidentiary material in 27 cases in just 30 hours.
The UK police aren’t alone. They are among the fast-growing cohort of law enforcement agencies employing new artificial intelligence technologies to help solve crimes.
In the summer of 2024, the Indianapolis Metropolitan Police Department brought in a team of AI researchers to go over their work on a 20-year-old unsolved murder. The group used an AI tool to create a timeline of events by analyzing 3,000 pages in the investigative file, including handwritten notes, the Indianapolis Star reported.
And in January 2024, a Miami Police Department official told lawmakers in a U.S. Senate subcommittee hearing that facial recognition technology and AI tools helped dramatically reduce homicide and violent crime rates in his city in recent years.
In addition to facial recognition software, police are using AI to pinpoint the location of gunshots in real time, narrow down areas populated with violent offenders with outstanding warrants and identify stores targeted by shoplifting rings, experts told A&E True Crime.
But as law enforcement agencies ramp up use of artificial intelligence to help solve crimes, experts warn the technology is still in its infancy and should only be used to augment traditional investigative methods.
“We have been using this technology for quite some time, but it is growing in a way that there is significant risk,” Adam Scott Wandt, an associate professor at John Jay College of Criminal Justice, tells A&E True Crime. “These are experimental tools, and we still need human analysis so things don’t go awry.”
Identifying Criminals and Their Patterns With AI
During his testimony before the U.S. Senate subcommittee, Miami Police Assistant Chief Armando Aguilar said the department treats facial recognition matches like an “anonymous tip,” according to a transcript of the proceedings.
“Most importantly, our policy emphasizes that facial recognition matches do not constitute probable cause to arrest,” Aguilar said during the hearing. “We establish a policy limiting who has access to our facial recognition platforms, and we disclose our facial recognition uses to defense counsel in criminal cases.”
Aside from facial recognition software, Miami Police, along withpolice departments in over 160 cities the U.S., use ShotSpotter, an AI powered program that pinpoints the location of gunshots in real time, Aguilar said. The department, in conjunction with the Bureau of Alcohol, Tobacco and Firearms, also has an in-house ballistics program and gun intelligence center powered by AI. (A number of cities—including Chicago, New Orleans and Atlanta—have also discontinued their use of ShotSpotter, claiming it’s not effective, costly and actually wastes police officers’ time.)
“The Miami Police Department has successfully leveraged artificial intelligence over the past few years to great effect,” Aguilar said during the hearing. “We use ballistic evidence to connect the dots between shootings and the violent actors that are victimizing our communities.”
Law enforcement agencies in other jurisdictions are experimenting with AI to prevent crimes before they occur. For example, Raleigh, North Carolina-based research organization RTI International collaborated with the Greensboro Police Department to develop an automated tool powered by AI to narrow down searches for violent fugitives.
“Law enforcement agencies often lack the tools or a standardized process to prioritize warrant service, with thousands of unserved warrants potentially overwhelming their ability to successfully carry out this function,” Jacob Cramer, an RTI senior policing researcher, tells A&E True Crime. “The tool helps officers determine the priority and location of active warrants and view these warrants on a map along with other data associated with the warrant such as [a suspect’s] full criminal history.”
An officer can prioritize which fugitives to target based on the historical likelihood of an individual committing a violent crime before their warrant is served, Cramer says.
Outside of North Carolina, RTI is conducting AI projects for two police departments in California to “combat organized retail theft and motor vehicle theft.” Cramer declined to identify the jurisdictions.
AI Poses Risks to Civil Liberties
As AI crime solving becomes more advanced, the technology could be used to provide forensic expertise about evidence, as well as scanning social media platforms for potential offenders before they commit crimes, says Wandt. For example, law enforcement agencies could possibly use artificial intelligence to identify and track pedophiles and terrorists, he says. But he acknowledges that utilizing technology in that manner could subject people to unconstitutional searches and seizures.
“There are significant legal challenges to overcome before it is available for law enforcement use,” Wandt says.
Divyansh Kaushik, an AI researcher and a former senior fellow at the Federation of American Scientists, tells A&E True Crime that AI tools used for facial recognition or “predictive policing” do not automatically take into consideration what constitutes probable cause for an arrest.
“The courts have generally been cautious in allowing AI-driven tools to replace traditional forms of evidence gathering,” Kaushik says. “Predictive policing may suggest that a crime is likely in a certain area, but this by itself cannot justify searches or arrests without independent corroboration through eyewitness testimony.”
Related Features:
Robot Cops Are Coming Soon. How Much Power Will They Have?
If Google Can Have Your Data, Can Police Investigating Crimes Have It Too?
