CS547 Human-Computer Interaction Seminar (Seminar on People, Computers, and Design)
Fridays 11:30am-12:30pm PT · Gates B3 · Open to the public|
J. Nathan Matias
Cornell University
Can we govern algorithms with science?
December 2, 2022
The governance of adaptive algorithms is one of society’s most pressing scientific challenges. In public discourse, these algorithms decide what we are allowed to say and rank the information we see. Because they adapt to human behaviors that they also influence, their actions have been hard to predict— representing a risk to society and a challenge for anyone who would govern their behavior. What kinds of knowledge can help us observe and govern these feedback loops between human and machine behavior? And how might this governance challenge require us to rethink how we design the software systems that support this research? This talk will summarize the dilemma of predicting, preventing, and intervening effectively on human-algorithm behavior, and how we might develop the knowledge needed to do so. |
|