There are good reasons to hold AI developers accountable for some of the negative outputs exhibited by their models, such as racially biased responses, “but in order to design effective mitigation strategies, we really need to think about regulatory and legal frameworks that focus on the ...companies actually using these models in real-world scenarios,” says Stanford Law’s Julian Nyarko, an AI expert and co-author of a new paper, “Breaking Down Bias: On The Limits of Generalizable Pruning Strategies.”

Working with Stanford Law research fellows Sibo Ma and Alejandro Salinas, along with Princeton computer science professor Peter Henderson, Nyarko probed deep into the internal mechanisms of LLMs to identify and mitigate the sources of biased outputs. https://stanford.io/3EzTaOS

Stanford Law Professors and constitutional law experts Jonathan Gienapp and Michael McConnell recently debated “History and Originalism: A Troubled Relationship,” at a Constitutional Law Center event. “Too often, originalists fail to take seriously the history they have pledged to follow,” ...Gienapp said. McConnell, a widely recognized proponent of originalism, praised Gienapp’s latest book, “Against Constitutional Originalism: A Historical Critique,” but said the title and some assertions were misleading. What Gienapp is ultimately arguing, McConnell said, is that originalists simply “are not originalist enough.”

Read the Legal Aggregate Q&A covering some of the high points of their conversation and watch a recording of the event. https://stanford.io/4gNVNdm

What is the Federal Vacancies Reform Act, and why is it significant? Stanford Law Professor Anne Joseph O’Connell, an expert in this area, is a guest on the Opening Arguments podcast to discuss how Trump's previous maneuvers raised questions about this law and what it means for future ...appointments.

Listen to the episode here: https://stanford.io/4gBFZtV