Prosecutors turn to AI for evidence management and analysis

As police departments, sheriffs and ultimately prosecutors’ offices investigate crimes, the nature of the evidence they rely on has changed dramatically.

No longer do law enforcement agencies use just fingerprints, witness statements and DNA. Now, they look to various forms of digital evidence, like footage from smartphones, dashcams, body-worn cameras, closed-circuit television and elsewhere. That means a massive influx of data to sift through, different ways evidence is collected and filed, and different file types that may be unsupported by some software systems.

For prosecutors’ offices that are already drowning in high caseloads and mountains of paperwork all amid a lack of staff and funding, this digital evidence could further hinder their ability to get those caseloads down and pursue justice. Prosecutors all agree it is a challenge.

“Our goal is the ethical pursuit of justice,” Randi Freese, the state’s attorney for McHenry County, Illinois, said in a statement. “We have a duty to seek the truth above all else, and that truth is often revealed through digital evidence. However, as our law enforcement partners adopt more digital systems, digital evidence is becoming increasingly difficult to collect, analyze, understand and share through conventional means.”

“With the increasing use of body-worn and in-car cameras, cell phones, video, and home surveillance, digital evidence is now a key component in almost every case,” Ryan McGuirk, chief investigator for the Monterey County, California, District Attorney’s Office, said in a statement.

But an increasing number are turning to artificial intelligence, machine learning, the cloud and other emerging technologies to get the problem under better control, alongside companies like NICE Justice, which announced partnerships with McHenry and Monterey Counties as well as Calcasieu Parish, Louisiana.

Evidence management is a crucial way that AI will be used by some prosecutors’ offices. Calcasieu Parish, for example, previously had to take digital evidence from a dozen public safety agencies, then manually download that evidence from hard drives, thumb drives and other physical media and upload it so it can be shared. Now, the parish has one portal through which those agencies can upload evidence, and discovery can be shared easily with defense attorneys.

“Prosecutors have to deal with many, many ways of collecting evidence,” Rod Guy, NICE’s vice president of strategy, said in an interview late last year. “The way they collect it from one agency is completely different from another, and it changes the way they have to process that evidence. The way they gather it from a sheriff is different from the local police department, and that just creates extra time and work. They literally have staff that do nothing but log in every morning and download things that come in from law enforcement.”

Such innovations are already having an impact, Guy said, explaining that having one unified platform for digital evidence management meant that a law enforcement agency could finally play a video that they had in their possession for years and were able to make progress on a homicide cold case by doing so.

Once they have the evidence in their possession, AI can help analyze it too. That analysis includes automated case building, video and audio transcription, optical character recognition to convert images of text into a machine-readable format, finding evidence connections, analytics and ever-controversial facial recognition.

Those analysis capabilities help “offload much of the legwork, costs and complexities of managing digital evidence,” Scott Hunsaker, director of IT for the Calcasieu Parish District Attorney’s Office, said in a statement. “This will allow our staff to get back to the work of pursuing justice and prosecuting cases.” 

Technology also makes it easier for prosecutors to retain evidence for possible appeals in the future, and easier for them to comply with states’ various statutes of limitations allowing for those appeals. 

“An appeal comes seven years into a 20-year retention policy, [prosecutors] have the ability to get access to [evidence], whereas today it might be in a box, in a warehouse somewhere and hard to get to,” Guy said.

While using AI and machine learning is a controversial approach in law enforcement, public safety and criminal justice, agencies and governments are slowly embracing it. The Los Angeles County Public Defender’s Office, for example, has digitized records and uses AI and automation to help provide clients with an effective defense. And as states are looking to automatically clear eligible criminal records, technology has a role to play in streamlining that process. Guy said technology holds much promise in this area.

“I think there are more opportunities in that space, helping [agencies] with even more efficiencies and maybe even decision support,” he said. “I wouldn't go so far as to say we'll help make decisions for them, but to find information and surface insights as there's lots of opportunity there.”