Artificial Intelligence: Law, Ethics, and Policy

Catalog description

Machine intelligence is rapidly improving. Recent surveys of top AI experts suggest there is around a 10% chance that by 2027 machines will be smarter than humans in nearly all tasks, and about a 50% chance of this happening by the 2040s. Forecasting models add that there may be a 50% chance of ?transformative AI? by 2033, systems capable of reshaping society as profoundly as the Industrial Revolution. The development of such generally intelligent systems will transform society, the economy, and global politics. The potential benefits are tremendous, but so are the potential harms, which some experts argue are greater than the risks from nuclear war and climate change. This course will examine the technical, political, and security challenges to the beneficial development of AI. How can we build, and prove to others that we built robust, beneficial AI? What legal and ethical problems are posed by autonomous machines? How will military conflict and strategy adapt to powerful AI-enabled drones, cyber weapons, military decision-making, and intelligence gathering? How can we avoid harmful AI arms races? Reducing existential risks from AI is a global public good: what global political institutions are best suited to providing it? (MPA-DA, MPA-DATSCI, MPA-SCITEC, MPA-TPA)