An AI tool used in thousands of criminal cases is facing legal challenges
Law enforcement agencies and prosecutors from Colorado to New York have turned to a little-known artificial intelligence tool in recent years to help investigate, charge and convict suspects accused of murder and other serious crimes.
Law enforcement agencies and prosecutors from Colorado to New York have turned to a little-known artificial intelligence tool in recent years to help investigate, charge and convict suspects accused of murder and other serious crimes.
But as the software, called Cybercheck, has spread, defense lawyers have questioned its accuracy and reliability. Its methodology is opaque, they’ve said, and it hasn’t been independently vetted.
The company behind the software has said the technology relies on machine learning to scour vast swaths of the web and gather “open source intelligence” — social media profiles, email addresses and other publicly available information — to help identify potential suspects’ physical locations and other details in homicides and human trafficking crimes, cold cases and manhunts.
The tool’s creator, Adam Mosher, has said that Cybercheck’s accuracy tops 90% and that it performs automated research that would take humans hundreds of hours to complete. By last year, the software had been used in nearly 8,000 cases spanning 40 states and nearly 300 agencies, according to a court decision that cited prosecutors in a New York case that relied on the tool.
In the New York case, a judge barred authorities from introducing Cybercheck evidence last year after having found that prosecutors hadn’t shown that it was reliable or well-accepted, the decision shows. In another ruling last year, an Ohio judge blocked a Cybercheck analysis when Mosher refused to disclose the software’s methodology.
Rating: 5