Berkeley AI Risk
Interdisciplinary discussion of risks to society posed by artificial intelligence
About Berkeley AI Risk
Berkeley AI Risk is an interdisciplinary community of faculty and students at UC Berkeley concerned about risks to society posed by artificial intelligence and how to mitigate them.
Scholars at Berkeley concerned about AI risk come from many units on campus, including:
Agricultural & Resource Economics
Ancient Greek & Roman Studies
Applied Science and Technology
Biostatistics
Center for Effective Global Action
Center for Science, Technology, Medicine & Society
Center for Information Technology Research in the Interest of Society
College of Computing, Data Science, and Society
Cognitive Science
Computational Precision Health
Computer Science
Criminal Law & Justice Center
Data Science
Demography
Economics
Education
Electrical Engineering and Computer Science
English
Environmental Science, Policy and Management
Epidemiology and Biostatistics
Goldman School of Public Policy
History
Industrial Engineering & Operations Research
Information and Data Science
Information Technology
Institute for Data Science
Interdisciplinary Studies
Labor Center
Law
Logic
Management
Mathematics
Music
Neuroscience
Philosophy
Political Science
Public Policy
School of Information
School of Law
Social Science Data Laboratory
Sociology
Statistics
Theater, Dance, and Performance Studies
Key Questions
- How do we prevent the gradual disempowerment of humans as more societal roles are turned over to AI systems?
- How do we avoid concentrations of power and rising inequality if the value of human labor declines as the value of AI capital increases?
- How might economic, geopolitical, and other selection pressures on AI systems threaten to erode the safeguards we attempt to put in place to ensure these systems are safe?
- Humans have collectively chosen not to develop some technologies, such as human cloning, and not to proliferate others, such as nuclear weapons; is it possible to do the same concerning autonomous, generally intelligent AI systems?
Speaker Series
Talks by experts from academia, civil society, and industry on AI risks and mitigations
Reading Group
Discussions of papers and research on AI risk, safety, ethics, and alignment
Organizers
Will Fithian (UCB Statistics)
Wes Holliday (UCB Philosophy and Logic)