UMass Law UMass Law: Professors Farber and Vyas present at Law Symposium on Data, AI, and the Constitution at University of Pennsylvania Law School

University of Massachusetts Dartmouth
UMass Law UMass Law: Professors Farber and Vyas present at Law Symposium on Data, AI, and the Constitution at University of Pennsylvania Law School
Professors Farber and Vyas present at Law Symposium on Data, AI, and the Constitution at University of Pennsylvania Law School

Professors Farber and Vyas presented their latest article titled “Truth and Technology: Deepfakes in Custodial Interrogations” at the University of Pennsylvania Journal of Constitutional Law Symposium

Professors Hillary Farber and Anoo Vyas

Professors Farber and Vyas presented their latest article titled “Truth and Technology: Deepfakes in Custodial Interrogationsat the University of Pennsylvania Journal of Constitutional Law Symposium on January 31, 2025. The Symposium focused on Data, Artificial Intelligence and the Constitution. The article explores the constitutionality of police use of generative AI in a custodial interrogation. Professor Farber presented their work on the first panel of the morning, AI and the Fourth Amendment: Deepfakes, Data & Dilemmas. Professor Vyas moderated the afternoon panel, Vindicating Rights Amid Algorithms: Constitutional Limits on Automated Decision Making.

In their co-authored article, Farber and Vyas explore the ethical and legal concerns surrounding the use of generative AI in police interrogations. Traditionally, courts have permitted law enforcement to use deceptive tactics in interrogation, like falsified forensic reports or fabricated witness statements, to extract confessions. With advancements in AI, however, police can now create convincing fake evidence, such as deepfaked videos or audio recordings, that could be used to pressure a suspect into confessing. Farber and Vyas raise concern that these types of ploys may well exceed the bounds of due process and exacerbate the potential for false confessions.

Furthermore, the article asserts that use of AI in interrogation tactics can be coercive and should not be directed at any suspect because these tactics disrupt our sense of justice in a civil society. Presently, few regulations exist to constrain how generative AI can be used in society at large. Farber and Vyas advocate for the construction of laws that apply to generative AI in civil society, and maintain that it is within reach to prescribe measures that protect the constitutional rights of suspects in interrogations.

The article is forthcoming in Volume 27 of the University of Pennsylvania Journal of Constitutional Law.