Stanford Seminar - AI Audits and Autonomy

The podcast features a presentation by a speaker detailing their work on algorithmic audits, usable control, and autonomy in sociotechnical systems. The speaker begins by discussing the history of algorithm regulation, using the Sabre system as an example, and then transitions to their work on housing audits using sock puppets to identify discrimination in online housing platforms and advertising. They explore different types of audits, including narrative visualization audits, and highlight the importance of awareness and control in user experiences. The speaker also presents research on control settings, placebo effects, and user expectations, emphasizing the need for transparency and contestability in AI systems. The presentation concludes with a discussion of community-based work and the importance of legal guidance in developing just and accountable sociotechnical systems, followed by a Q&A session with audience members.