The Wadhwani AI Center has relaunched its AI Policy Podcast and released a new commentary exploring renewed interest in licensing powerful AI models, signaling a significant focus on AI governance and policy shifts.
Director Aalok Mehta hosted the first episode of the relaunched podcast with Matt Mande, where they discussed Pope Leo XIV's recent encyclical on AI and President Donald J Trump's postponed executive order concerning AI and cybersecurity. New episodes are slated for release every two weeks, with a series of guest interviews planned for the coming months.
Mehta's commentary, titled "AI Licensing Comes Full Circle," delves into the evolving AI policy discourse. It highlights a shift from a 2023 focus on safety to current discussions on AI adoption and buildout, noting that Anthropic's Mythos model has prompted U.S. policymakers to reconsider a light-touch regulatory approach, echoing OpenAI CEO Sam Altman's earlier calls for licensing.
Should policymakers advance an AI model licensing regime, Mehta recommends three key actions: tailoring the regulatory framework to AI's unique characteristics, investing in evaluations that accurately reflect real-world performance, and engaging in robust post-market oversight.