Regulating Artificial Intelligence

Even just a generation ago, interest in "artificial intelligence" (AI) was largely confined to academic computer science, philosophy, engineering research and development efforts, and science fiction. Today the term is widely understood to encompass not only long-term efforts to simulate the kind of general intelligence humans reflect, but also fast-evolving technologies (such as elaborate convolutional neural networks leveraging vast amounts of data) increasingly affecting finance, transportation, health care, national security, advertising and social media, and a variety of other fields. Conceived for students with interest in law, business, public policy, design, and ethics, this highly interactive course surveys current and emerging legal and policy problems related to how law structures humanity's relationship to artificially-constructed intelligence. To deepen students' understanding of current and medium-term problems in this area, the course explores definitions and foundational concepts associated with "artificial intelligence," likely directions for the evolution of AI, and different types of legally-relevant concerns raised by those developments and by the use of existing versions of AI. We will consider distinct settings where regulation of AI is emerging as a challenge or topic of interest, including autonomous vehicles, autonomous weapons, AI in social media/communications platforms, and systemic AI safety problems; doctrines and legal provisions relevant to the development, control, and deployment of AI such as the European Union's General Data Protection Regulation; the connection between the legal treatment of manufactured intelligence and related bodies of existing law, such as administrative law, torts, constitutional principles, criminal justice, and international law; and new legal arrangements that could affect the development and use of AI. We will also cover topics associated with the development and design of AI as they relate to the legal system, such as measuring algorithmic bias and explainability of AI models. Cross-cutting themes will include: how law affects the way important societal decisions are justified, the balance of power and responsibility between humans and machines in different settings, the incorporation of multiple values into AI decision making frameworks, the interplay of norms and formal law, the technical complexities that may arise as society scales deployment of AI systems, and similarities and differences to other domains of human activity raising regulatory trade-offs and affected by technological change. Note: The course is designed both for students who want a survey of the field and lack any technical knowledge, as well as for students who want to gain tools and ideas to deepen their existing interest or background in the topic. Students with longer-term interest in or experience with the subject are welcome to do a more technically-oriented paper or project in connection with this class. But technical knowledge or familiarity with AI is not a prerequisite, as various optional readings and some in-class material will help provide necessary background. Requirements: The course involves a mix of lectures, in-class activities, and student-led discussion and presentations. Requirements include attendance, participation in planning and conducting at least one student-led group presentation or discussion, two short 3-5 pp. response papers for other class sessions, and either an exam or a 25-30 pp. research paper. After the term begins, students accepted into the course can transfer, with consent of the instructor, from section (01) into section (02), which meets the R requirement. CONSENT APPLICATION: We will try to accommodate as many people as possible with interest in the course. But to facilitate planning and confirm your level of interest, please fill out an application (available at https://bit.ly/2MJIem9) by TBA. Applications received after the deadline will be considered on a rolling basis if space is available. The application is also available on the SLS website (Click Courses at the bottom of the homepage and then click Consent of Instructor Forms).