The University of Wisconsin Holtz Center for Science and Technology studies hosted a panel which discussed how both big data algorithms and individual actor decisions in the criminal justice system can lead to discriminatory practices.

The panel consisted of University of California-Irvine professor criminology Simon A Cole and Washington and Lee University law professor Margaret Hu. The panel was moderated by University of Wisconsin associate law professor Cecelia Klingele, who began the panel discussion by speaking about surveillance in the criminal justice system.

“The collection of data in Criminal Justice in the U.S is complicated by many factors. One problem is the fragmentation of data from criminal justice agencies,” Klingele said. “That fragmentation of data and the lack of controls leads us to the inability to gather aggregate data about the functioning of our criminal justice agencies, and the way they are either positively or negatively affecting our communities.”

Klingele said the criminal justice system is primarily governed at the discretion of powerful individual actors. But, the tough decisions these actors make could change remarkably as algorithms begin to dominate decision making.

As AI programs proliferate, researchers investigate implications of machine-driven decisionsFrom facial recognition to self-driving cars, job hiring algorithms to online chatbots, targeted advertising to Google Translate, artificial intelligence programs Read…

Hu delivered a presentation on big data and cyberspace in the criminal justice system.

Hu said the government uses mass surveillance to gather data on people. People are “flagged” based on what they do on the internet and when they do it, Hu said.

“In a big data, cyber-surveillance world, the analytical method is to start with the data,” Hu said. “The data becomes suspicions — it’s not necessarily that people are suspicious anymore. This leads to pre-crime ambitions — to stop crime before it starts.”

Hu talked about how recent changes in mass surveillance have marked a major transformational moment in criminal justice history.

As of 2000, 25 percent of all the world’s information was digitized. In 2012, 98 percent was digitized. Experts predict the amount of data we digitize will double every two years by 2020, at which point there will be 500,200 GB of data for every person on earth, Hu said.

“Extreme vetting and forms of data collection are going to eventually influence every right, privilege freedom and liberty that we have. We’re already at the earliest stages of it,” Hu said.

UW experts weigh in on use of defendant-risk assessment tool by Wisconsin judgesThe Wisconsin Supreme Court may review the constitutionality of a tool used by state judges to help determine the level of Read…

Cole presented on the history of how the criminal justice system has identified criminals.

After talking about the discovery of fingerprinting, Cole discussed how race and individualization could factor into DNA and mass surveillance today.

“DNA brings about a kind of public policy problem: How big should the DNA database be, and who should be in it?” Cole said.

Cole recognized how important it is to keep the DNA pool “fair” by containing many people, but said the way people get into the system doesn’t always allow for fairness. He said arrests are usually not “neutral,” as they almost always allow for significant discretion to individual actors in the criminal justice system.

But, Cole said algorithms take away all individual discretion in the system, and could possibly target people who are not criminals.

Cole arrived at two solutions to keep the criminal justice system “fair”. He said to either include everything about everyone in the DNA database, or focus more on actual behavior — not a prediction based on algorithms.

“With criminal justice algorithms, we are just predicting who will be a criminal and who looks dangerous in these algorithms, which seems like it will be just as discriminatory in the end,” Cole said.