Duke Experts Talk Artificial Intelligence With Congressional Staff

Vincent Conitzer talks with congressional staff on Feb. 15.
Vincent Conitzer and other Duke faculty told congressional staff that AI presents lawmakers with unusual regulatory challenges.

Increased federal funding and ethical inquiry are needed to best develop America’s artificial intelligence capabilities, argued three Duke experts in a congressional briefing on Capitol Hill on Feb. 15.

Duke Professors Vincent Conitzer, Nita Farahany and Walter Sinnott-Armstrong spoke broadly about the ethical implications of what the advent of A.I. means for medicine, lethal weapons, automobiles and unemployment. Conitzer’s presentation offered a definition of A.I., Sinnott-Armstrong explored the ethics of lethal autonomous weapons and Farahany dove into the legal questions A.I. will challenge.

Artificially intelligent systems already excel at games of probability and prediction but fail at games of context and interpretation, said Conitzer in his presentation.

This program began a three-part Duke in DC series for congressional staff exploring policy implications for human-A.I. collaboration. The second session will focus on transportation and the third will focus on health care.

Roughly 70 congressional staff joined the lunchtime presentations. Attendees asked a variety of questions on the pace of A.I. development and the importance of technological competition in the international space.

Several attendees asked about the hazards of racial bias in data analyses and some asked about the ethics of bioengineered organs.

Several questions focused on how to regulate without stanching innovation. Farahany offered the idea that regulation was not a before or after phenomenon, but rather a parallel process between innovators and the government.