NACDL - Challenging the Use of Secret Algorithms and Technologies in Criminal Cases

Challenging the Use of Secret Algorithms and Technologies in Criminal Cases

This program guided defense attorneys through the challenges of obtaining “black box” evidence and confronting the flaws, bugs, and biases embedded in technologies deployed in the criminal justice system.

Unlocking the Black Box: Challenging the Use of Secret Algorithms and Technologies in Criminal Cases.

Increasingly, technology is being used in the criminal legal system to identify DNA samples, assess risk factors in release decisions, match images in face recognition technology and even determine where or who the police should be patrolling and investigating. From probabilistic genotyping to “risk assessment” software, the recent explosion of emerging technologies has transformed almost every aspect of the criminal legal system. Increasing amounts of data and evidence are being interrogated and generated using software systems that are kept from defense teams, the courts, and the general public. The assertion of “trade secrets” by the companies who develop these tools deprive defense lawyers of access to information on how the software was constructed and the opportunity to assess its accuracy, credibility and reliability. How can criminal defense attorneys understand and confront the limitations of software-based evidence and machine learning algorithms in criminal proceedings?

This webinar from February 17, 2021 featured Megan Graham, Clinical Supervising Attorney in the Samuelson Law, Technology & Public Policy Clinic at UC Berkeley, School of Law; Joanna Kloet, Assistant Federal Defender in the Western District of Michigan; and Nitin Kohli, PhD candidate at the UC Berkeley School of Information.