Rebooting Democracy in the Age of AI

A new paper introducing the UbuntuGuard benchmark questions whether strong results on English-language safety tests consistently translate into responsible use. Built from policies developed by 155 African domain experts across ten languages and six countries, UbuntuGuard's framework assesses whether AI tools comply with the norms that shape services in non-Western contexts. The findings suggest that institutions, wherever they operate, need the capacity to define their own standards before using these tools to improve public-sector outcomes.

Max Stier argues that rebuilding trust in government begins inside agencies by recognizing and elevating the everyday excellence of career civil servants. Drawing on the Partnership for Public Service and federal workforce data, he argues that internal culture, not just external messaging, shapes public confidence. The post explains how AI can cut paperwork, reduce bias, recognize good work in real time, and tailor praise, helping leaders boost morale, reward strong performance, and better serve the public.

This week on Reboot Democracy, Beth Simone Noveck and Dane Gambrell explain why more people need to learn how to "do engagement" in the AI era and invite co-creation of a new online course. Sofía Bosch Gómez, Joanna French, and Belén Farmer Martinez reflect on co-designing AI with families in San Francisco. Merve Hickok examines South Korea’s investments in small models. Beyond Reboot, federal AI deployments have nearly doubled, with the DOT planning to use AI to write regulations; labor leaders are pressing for stronger worker protections; and fights over data centers and digital public infrastructure highlight how power and accountability are being negotiated in real time.