Researchers are calling for open and free access to U.S. court records and building an AI tool to analyze them.
Why it matters: Court records are publicly available but expensive to access and difficult to navigate. Freeing up that data — and using machine learning tools to make sense of it — would help make the justice system more just.
While records for Congress and executive agencies are free on the internet, federal courts charge $0.10 per printed page to view any record online.
- That makes it difficult and costly for researchers, journalists and ordinary citizens to tap the raw data needed to understand the inner workings of the U.S. justice system.
What’s new: In one example of the kind of analysis that could be possible with open access, researchers from Northwestern University used an algorithm to scan court records and determine how often judges granted waivers for the $400 fee required to file a federal lawsuit.
- While there is no uniform standard for granting waivers, the researchers found unexpectedly huge variations. In one district, the approval rate varied from less than 20% for some judges to more than 80% for others.
- With open access “we can get a fuller picture of what the systematic trends are and make it all easily accessible,” says Adam Pah of Northwestern’s Kellogg School of Management and Organizations and one of the co-authors of the study.
What’s next: The Northwestern researchers are working on an AI-powered platform called SCALES-OKN that would make federal courtroom data accessible to the public and easily analyzable, linking data in the courts to information outside them.
- Such a platform could be a potent tool for uncovering hidden bias over money or race in the justice system, says Pah.
The big picture: AI is already being used in the criminal justice system for policing and sentencing, but experts say it too often perpetuates a biased system. Unleashing AI on open court records could provide a welcome opportunity to use technology to further justice, not curtail it.
Go deeper: Coronavirus accelerates AI in health care