Policy Unpacked: Can Governments Keep up with AI?

Why It Matters

Adopting AI could materially raise government capacity and citizen outcomes, but without clear policies and safeguards the risks to equity, privacy and public trust are significant. Effective governance will determine whether AI’s public-sector benefits are realized or undermined.

Summary

At Oxford’s Blatnik School of Government, experts argued that while AI is dual-use and poses risks, governments should adopt specialized AI tools to improve public services. They highlighted concrete wins in healthcare (automated scans), agriculture (Plantics and ISAT crop and climate tools), literacy, energy-grid optimization, traffic management (Glide), and routine administrative tasks like transcription and translation. AI’s core advantages are scale, speed, efficiency and customization, enabling governments to deliver services at volumes and precision previously unattainable. The conversation emphasized managing harms through systematic governance rather than avoiding the technology outright.

Original Description

Can governments keep up with AI? Fellow of Practice Dr Aaron Maniam and Master of Public Policy student Aliénor Rougeot explore one of today’s biggest policy questions in the second episode of our Policy Unpacked series.

In this new series from the Blavatnik School of Government, professors and students sit down together to unpack the issues shaping today’s world.

Aaron Maniam is Fellow of Practice and Director of Digital Transformation Education. Previously a policymaker in the Singapore government, his work now focuses on issues connecting technology, public policy and public administration.

Aliénor is a Canadian and French climate justice advocate with experience in both national and grassroots civil society organisations. During her time on the MPP, she has been exploring how to make public policy more responsive to the pace of technological change.

Chapters:

00:00–00:52 Introduction

00:53–07:07 Why should governments use AI?

07:08–09:09 Examples of governments using AI chatbots

09:10–10:15 Governments leading in the AI space

10:16–12:43 Cultural differences across AI usage

12:44–19:06 The risks: Replication of biases

19:07–20:07 Can we trust AI’s outputs?

20:08–21:44 The responsibility of tech companies

21:45–25:15 Who is accountable when governments use AI

25:16–30:09 The tech skills government officials need

30:10–35:28 The risks: using AI for mass surveillance

35:29–40:38 What governments need to get right with AI

Blavatnik School of Government,

University of Oxford

http://www.bsg.ox.ac.uk/