Coursera Flash Sale
40% Off Coursera Plus for 3 Months!
Grab it
Explore two critical challenges in AI governance through this conference talk featuring Simons Institute Law and Society Fellows Rui-Jie Yew and Greg Demirchyan. Examine the first alignment problem: the lack of thorough understanding of AI models that makes risk identification difficult, coupled with potentially unreliable auditing tools that may not provide sufficient assurances of model safety and alignment. Learn how establishing the faithfulness of AI explanations remains challenging at scale, creating regulatory difficulties for current governance proposals and exposing misalignment between proposed interventions and their feasibility. Investigate the second challenge involving unintended effects of AI system design, deployment, and framing to minimize regulatory costs, where safety-enhancing technologies like privacy-preserving methods and AI evaluations simultaneously shape regulatory oversight terms and may be deployed in ways that conflict with regulatory goals. Discover approaches to governance that attempt to reduce these tensions by designing adaptable regulatory systems that can evolve as understanding of transformative AI technology improves, and explore steps toward robust oversight of AI systems. Gain insights from Yew's research at the intersection of artificial intelligence, policy, and law, including her work recognized at FAccT, AIES, and ACM Symposium on Computer Science and Law, as well as Demirchyan's expertise in technology law, AI governance consulting, and algorithmic fairness research developed through his roles as CEO of Fairlogic and former general counsel for tech startups.
