AI and Public Policy

What an AI and public policy degree actually is

Someone has to write the rules for artificial intelligence — decide what an audit of a frontier model looks like, when a government may use facial recognition, how compute exports get controlled. AI policy is the field staffing those decisions, and it is hiring faster than any other policy specialisation this decade: regulators, safety institutes, standards bodies and the policy teams of the AI labs themselves are all recruiting from a talent pool that barely existed five years ago.

Because the field is new, no two universities package it under the same name — which is exactly why a name map matters here. Cambridge teaches it as an ethics MPhil inside its technology-and-humanity institute; Oxford routes it through the Oxford Internet Institute's social-science MSc; UCL embeds it in an engineering-facing public-policy degree (STEaPP); Sussex's SPRU approaches from innovation policy; and Berlin's Hertie School teaches the data science itself alongside the policy. Searching "AI policy masters" will surface almost none of these.

What you study — and the quantitative bar

Three distinct entry lenses hide under one label. The ethics lens (Cambridge's MPhil in Ethics of AI, Data and Algorithms) is essay-based: philosophy of technology, algorithmic fairness, a dissertation. The social-science lens (Oxford's MSc in Social Science of the Internet) mixes methods training with electives like AI for policymaking and digital ethics. The technical-policy lens is the most quantitative: UCL's STEaPP teaches policy engineering with real public bodies, and Hertie's two-year Master of Data Science for Public Policy has you writing machine-learning code — statistical programming is the core curriculum, not an elective.

Pick the lens by the bar you can clear and the credibility you need. The ethics and social-science routes ask for strong writing and a first degree in any rigorous discipline; the Hertie route expects mathematical maturity and rewards it with the rare double competence — the person in the room who has both trained a model and drafted a regulation.

Where it leads

AI governance units inside governments and regulators, the new national AI safety institutes, international processes (OECD AI, UN advisory bodies, EU AI Act enforcement), standards organisations, tech-policy think tanks, and the fast-growing policy and governance teams at AI labs and platform companies. The field's defining career fact: demand for people who can read both a model card and a statute vastly exceeds supply, and the shortage is worst inside governments — which is precisely the gap scholarship schemes exist to fill.

Who it suits — and who it does not

Where to study it: the programme map

Five verified programmes across three lenses — ethics (Cambridge), social science (Oxford OII), technical policy (UCL, Hertie) and innovation policy (Sussex SPRU) — running 9 to 24 months. Names differ so much that none contain the phrase "AI policy": search by the official titles below.

Every row verified against the official programme page; oldest verification 15 Jul 2026. Nothing here is a paid placement.

Application strategy and funding routes

Choose by lens, then by format. Cambridge's MPhil (9 months) and Oxford's OII MSc (10 months) are the prestige short routes — both fiercely competitive, both with autumn/winter deadlines that close earlier than typical UK programmes. UCL's STEaPP suits engineers who want London's policy market at the door; Sussex's SPRU frames AI inside the broader innovation-policy toolkit (covered in depth in our Science and Technology Policy guide); Hertie's two-year Berlin programme is the deepest technical build and the strongest EU-institutions springboard.

All four UK options fit Chevening's 9–12-month rule, and "helping my country govern AI" is currently one of the freshest, least-saturated scholarship narratives available — assessors have read a thousand generic policy essays and very few credible AI-governance ones. Hertie sits outside Chevening but inside DAAD territory and offers its own partial scholarships.

Frequently asked questions

Related fields