12/7 IER Seminar Series: Steve Dorton – “Trust Dynamics with AI in High-Consequence Work Systems”

Trust Dynamics with AI in High-Consequence Work Systems

Wednesday, 12/7/2022 from 1 – 2 pm

ISEC 532 & Zoom
Zoom: https://northeastern.zoom.us/j/96750528112?pwd=Z3M0b1Z0QTZFaWM3QzZ5bC92SjFUZz09

Steve Dorton

Principal Scientist for Sensemaking, Decision Making, and AI

The MITRE Corporation

Abstract:

Artificial Intelligence (AI) is often viewed as the means by which intelligence analysts will cope with the ever-increasing deluge of data from various sources. The best AI is moot, however, if analysts cannot trust the outputs of the AI to inform high-consequence decision making. A naturalistic study was performed to understand how intelligence professionals gain and lose trust in AI “in the wild.” The study assessed various trust factors proposed in the literature and identified various themes from interviews with intelligence professionals. We will discuss how to apply these findings to engineer more trustworthy AI for high-consequence decision applications.

Bio:

Steve Dorton is a Principal Scientist for Sensemaking, Decision Making, and AI at the MITRE Corporation. His research generally falls at the intersection of the social and computational sciences, focusing on how intelligent systems can help and harm human cognition in national security contexts. He also holds an adjunct lecturer appointment in the University of Maryland School of Public Policy, where he teaches social, ethical, and policy considerations for AI and big data.