How can you tell if an algorithm is any good? For University of Michigan assistant professor of information and public policy Ben Green, the answer lies less in accuracy scores, fairness metrics or any other technical measure than in what the system actually does once it’s out in the world. In his new book, Algorithmic Realism: Data Science Practices to Promote Social Justice (MIT Press), he offers data scientists a method for scoping, building and evaluating algorithms in the service of social justice.

In this episode, Tech Policy Press deputy editor Ramsha Jahangir examines the latest developments in Europe’s AI policy debate, from the push for safe, trusted and resilient AI to renewed calls for legislation on AI liability.

Risto Uuk, head of European policy and research at the Future of Life Institute, joins the podcast to discuss the emerging risks and the broader policy landscape. Then, Ramsha interviews Brando Benifei, a Member of the European Parliament, about the gaps left by the withdrawal of the proposed AI Liability Directive, the challenges of assigning responsibility across the AI value chain, and the case for a renewed European approach.

Pennsylvania offers a particular case study in which to consider the political battle over data centers that is playing out across the United States and beyond. A new report from Data & Society, "The AI Factory: Data Centers, Power, and Resistance in Late Industrial Pennsylvania," draws on a year and a half of fieldwork across the state. It examines how boosters sell data centers by invoking Pennsylvania's industrial past, and why so many communities aren't buying the promises made by industry and politicians. Justin Hendrix spoke with Data & Society’s Maia Woluchem and Ana Carolina de Assis Nunes and People’s Tech Project’s Hannah Sassaman about its findings.