STS10SI

As we delegate more and more societal responsibilities to Artificial Intelligence, we raise pressing ethical questions about what will happen if these systems aren't aligned with our values. Increasingly many AI experts across academia and industry believe there is an urgent need for both technical and societal progress across AI alignment, ethics, and governance to understand and mitigate risks from advanced AI systems and ensure that their contributions benefit humanity and the world. Intro to AI Alignment explores these questions in lectures and small discussion-based environments led by student facilitators with targeted readings, weekly quizzes and group discussions, and a small final project. After recapping recent advancements in AI development, we will start by exploring two sides of the AI alignment problem that prevent us from building AI systems that reliably understand and follow human-compatible values. Next, we'll discuss current harms from AI as well as risks that future systems could pose and arguments for and against the importance of various AI safety work. Finally, we will learn about existing AI safety technical research, efforts to implement policy and governance measures that reduce AI risk, and how you can personally contribute to AI safety. Basic knowledge about machine learning helps but is not required. Enrollment is by application only. View the full syllabus and apply online at https://linktr.ee/stanfordaialignment by Sunday, Dec 17, 2023 at 9:00 PM PST.