Almost everyone planning around AI regulation in the UK carries the same mental map: America deregulates and races for innovation, the European Union regulates and demands trust, China builds under state direction, and every other country is a variant placed somewhere between those three poles. It is a tidy map. It is also, on the evidence of what governments actually write in their AI strategies, a misleading one. A text-based analysis of 56 national strategies finds the UK sitting in a growth-and-delivery cluster that contains America and several EU members — but also Bangladesh, Brazil and Ethiopia. If you are using the three-model map to predict where UK rules are heading, you are reading the wrong chart.
The map most leaders are using is the wrong shape
The shorthand is everywhere in boardrooms and policy notes: pick your pole. Treat the US as the innovation-led model, the EU as the regulation-led model, China as the state-led model, and position everyone else by proximity. The trouble, as researchers Elise Antoine and Adam Chalmers argue in LSE Business Review, is that the three labels do not even describe the same kind of thing. “American innovation” is a claim about what AI should deliver. “European regulation” is a claim about how AI should be governed. “Chinese state-led development” bundles a governance style and a development agenda together. Three reference points, three different questions — collapsed into one axis.
That category error matters for anyone making decisions on the strength of it. A procurement lead who assumes the UK will track the American pole is predicting light-touch rules. A compliance team that files the UK next to Brussels is budgeting for an EU-style regime. Both are extrapolating from a map that was never built to bear the weight.
Strategic Reality: The US-EU-China model is a description of three countries, not a coordinate system. Using it to anticipate a fourth country’s regulatory direction assumes every national strategy is a blend of three fixed recipes. National strategies are not built that way, and the UK is a clear case of why the assumption fails.
The fix is to separate the two questions the bloc story conflates. How should AI be governed — with a lean towards ethics and regulation, or towards growth and innovation? And what should AI be used to achieve — how heavily does the strategy stress delivery themes such as infrastructure, public services and innovation deployment? Score national strategies along those two independent dimensions and the neat three-pole map dissolves into something more useful.
The real story: what 56 strategies actually say
The analysis drew on x.Machina, an open-access governance-mapping platform that catalogues flagship national AI strategies. The researchers took the most recent strategy document for each of 56 countries and scored it across eight themes: ethics, regulation, growth, innovation, public services, data, workforce and climate. The scoring is text-based rather than keyword-based, so two countries can land close together even when they use entirely different vocabulary — a meaningful methodological choice, because it measures emphasis rather than buzzword frequency.
From the eight themes the researchers built two composite measures. Governance orientation captures the balance between ethics-and-regulation language and growth-and-innovation language. Functional emphasis captures how strongly the text stresses delivery — growth, innovation, public services and data infrastructure. Clustering the 56 strategies along those two axes produced four archetypes, and none of them behaves the way the bloc map predicts.
| Measure | Definition | Why it matters for planning |
|---|---|---|
| Strategies analysed | 56 flagship national AI strategies | The comparison base — broad enough to test the three-model claim |
| Themes scored | Ethics, regulation, growth, innovation, public services, data, workforce, climate | The eight ingredients every strategy combines differently |
| Governance orientation | Ethics-and-regulation vs growth-and-innovation lean | Predicts regulatory direction better than a bloc label |
| Functional emphasis | Weight on delivery: infrastructure, public services, deployment | Predicts public-sector AI priorities and where capability concentrates |
| Archetypes that emerged | Four clusters, none anchored by the US, EU or China | The three-model map has no equivalent in the texts |
The headline finding is blunt: the anchors do not anchor. America does not sit at the head of an Anglosphere innovation bloc. The European Union does not occupy a single cluster. No identifiable “China model” grouping appears in the texts at all.
Critical Context: The four archetypes are Balanced (Ethics + Regulation), Growth-oriented (Climate + Data), Balanced (Economic + Public) and a small Low-salience cluster. Membership cuts across every political, economic and regional line the bloc story relies on. Germany shares a cluster with Saudi Arabia. America shares one with Bangladesh and Brazil. China shares one with India and Japan — two countries no analyst would call followers of a Chinese model.
What’s really happening: the UK is a delivery-led economy on paper
Here is the part that should reorganise UK planning. Britain does not sit in the Balanced (Ethics + Regulation) cluster where France and Germany sit. It sits in the Growth-oriented (Climate + Data) cluster — the same archetype as the United States, alongside Australia and Ireland, several EU members including Austria, Belgium, Denmark, Italy and Sweden, and a varied group of middle- and lower-income economies such as Bangladesh, Brazil, Colombia, Ethiopia and Moldova.
Read carefully, that placement says two distinct things. First, it partly confirms the UK’s own rhetoric: successive governments have positioned British AI policy as pro-innovation and delivery-focused, and the strategy text bears that out — the document leans towards growth and data infrastructure rather than ethics-and-regulation framing. Second, and more awkwardly, it demolishes the comfortable idea that the UK occupies a distinctive “third way” between American deregulation and European caution. On the evidence of the text, the UK is not bridging two poles. It is sitting squarely in a delivery-oriented group whose other members range from a G7 superpower to economies at a very different stage of development.
Reality Check: The “third way” positioning is a story the UK tells about itself, not a pattern the strategy text supports. What the document actually emphasises — growth, data, deployment, light governance framing — places Britain in company that has nothing to do with shared institutions or income levels. Strategic similarity here is about priorities, not peers.
The cross-cutting nature of the clusters is the deeper lesson. The countries that share the UK’s archetype are a mix politically, economically and regionally. That is precisely what makes the bloc map dangerous: it predicts that the UK will converge with its political allies, when the text shows it converging with whoever happens to weight delivery themes the same way — a set defined by emphasis, not by alliance.
One finding crystallises the point. China and Scotland — whose recent strategy is also in the dataset — sit in different archetypes and differ on governance orientation, exactly as a bloc-map reader would expect. Yet on functional emphasis, on how heavily each text stresses infrastructure, public services and deployment, the two are strikingly close. Two jurisdictions cast as opposites in the geopolitical story weight delivery almost identically. The similarities that matter for cross-border planning sit inside the archetypes, not between the famous poles.
The stakeholder picture: who has been reading the wrong chart
The practical cost of the three-model map falls unevenly. It misleads different decision-makers in different directions, and naming those directions is the first step to correcting course.
| Who | What the bloc map tells them | What the text analysis tells them instead |
|---|---|---|
| Compliance and legal teams | The UK will drift towards either US or EU rules | The UK strategy leans growth-and-delivery — closer to the US end on governance orientation, but not bound to either statute book |
| Public-sector AI buyers | UK priorities mirror its political allies | UK delivery emphasis matches a cross-cutting group; useful comparators are functional, not geopolitical |
| Multinationals mapping jurisdictions | Three regimes to plan for, everyone else a variant | At least four distinct emphasis patterns; “where does this country actually lean” beats “which pole is it near” |
| Investors and market entrants | Anglosphere markets cluster together | Britain, Australia and Ireland share an archetype with Brazil and Ethiopia — shared priorities, very different markets |
| Policy and public-affairs teams | The UK occupies a distinctive middle ground | The strategy text shows no distinctive third way; the positioning is rhetorical |
The throughline is that proximity on a bloc map is a poor proxy for the things businesses actually need to anticipate: regulatory direction, public-sector AI priorities and where institutional architectures are likely to converge. Antoine and Chalmers put it plainly — the three-model story will systematically mislead anyone using it to forecast those outcomes. Reading the strategy texts along the two dimensions is the more reliable basis.
Strategic recommendations: how to read national strategies properly
For UK organisations operating across jurisdictions, the correction is concrete and inexpensive. It is mostly a matter of replacing a lazy map with a sharper one.
- Retire “which pole is it near” as a planning question. Replace it with two: how does this country’s strategy lean on governance, and how heavily does it stress delivery? Those two answers predict more than any bloc label.
- Build functional comparators, not geopolitical ones. If you want to anticipate UK public-sector AI priorities, study the strategies that share its functional emphasis — not the ones that share its alliances. The useful neighbours may surprise you.
- Check which document you are reading. A country’s position depends on which text is treated as representative. The US analysis used the 2025 Action Plan, which places America in the growth-oriented cluster; the thinner March 2026 legislative recommendations would place it elsewhere. Always confirm you are comparing like with like.
- Treat the UK’s “third way” framing as marketing, not forecast. When the rhetorical positioning and the strategy text diverge, plan from the text.
Implementation Note: None of this requires proprietary intelligence. The x.Machina dataset and the analysis behind it are publicly available at nomosmachina.org. A compliance or public-affairs team can read the source strategies directly and score them on the same two dimensions, rather than inheriting a bloc map second-hand from commentary.
For smaller UK organisations, there is a quieter advantage here. The bloc map is the default in large consultancies and incumbent advisory functions; reading the actual texts is cheap, and doing it well produces a more accurate read of regulatory direction than the expensive shorthand everyone else is buying.
SME Advantage: A two-dimensional read of the relevant national strategies costs an afternoon, not a retainer. For a UK SME deciding which jurisdictions to enter or which compliance posture to adopt, that read is both more accurate and more affordable than the three-model briefing it replaces.
Hidden challenges in dropping the bloc map
Trading the three-model map for a text-based read is the right move, but it is not free of traps. Four are worth naming before you rely on the new approach.
The first is recency drift. A national strategy is a snapshot, and governments revise. The same country can move between archetypes as it publishes, amends and supersedes documents — the US Action Plan versus its later legislative recommendations is the clearest example. Any cross-jurisdictional read needs a date stamp and a refresh schedule, or it quietly goes stale.
The second is that emphasis is not enforcement. A strategy text tells you what a government says it prioritises, not what it has legislated or funded. The UK’s growth-oriented emphasis is a statement of intent; it does not by itself tell you what the eventual statutory regime will contain. Treat the text analysis as a guide to direction, then verify against actual legislation and spending.
Warning ⚠️: Do not over-read a single archetype. The four clusters are a coarse summary; genuinely important similarities — like the China-Scotland match on functional emphasis — sit inside the archetypes and need closer comparison than the four-way grouping can give. The map is a starting point for analysis, not a substitute for it.
The third is selection of the representative document. Many countries produce more than one AI policy text, and choosing which one stands for the country shapes the result. Be explicit about which document you are scoring and why, especially when a country has a comprehensive strategy and a thinner, more targeted legislative text that point in different directions.
The fourth is the temptation to build a new tidy story to replace the old one. The honest finding is that the field is varied and moving, not that there are secretly four blocs instead of three. The value is in reading specific countries on specific dimensions for specific decisions — not in swapping one oversimplification for another.
The strategic takeaway
The three-model map persists because it is easy to carry and easy to repeat. Its weakness is that it answers a question — which pole is this country near — that does not predict the things UK organisations actually need to know. The text of 56 national strategies shows a picture that is richer, more cross-cutting and more changeable than three poles can capture, with the UK sitting in a growth-and-delivery cluster that has more to do with shared priorities than shared politics.
Three points are worth carrying out of this:
- The UK reads as delivery-led, not as a third way. Its strategy text leans towards growth, data and deployment, grouping it with the US and a varied set of other economies — not bridging the US and EU poles.
- Functional comparators beat geopolitical ones. To anticipate UK regulatory direction or public-sector priorities, study the countries that share its emphasis, not the ones that share its alliances.
- The map is a snapshot, so date it and refresh it. Strategies are rewritten; positions move; always confirm which document you are reading before you draw a conclusion from it.
The next step for any UK team that has been planning around the bloc map is small and worth doing this quarter: pull the relevant national strategies, score them on governance orientation and functional emphasis, and compare the result against the assumptions in your current jurisdiction plan. Where they diverge, trust the text.
If you want to discuss what this means for your organisation’s cross-border AI planning, get in touch — or explore our other strategic insights on UK AI governance.
Analysis based on “How countries write their AI strategies — mapping the many models of governance” by Elise Antoine and Adam Chalmers, published in LSE Business Review on 28 May 2026. The underlying x.Machina dataset is publicly available at nomosmachina.org. This article is original Resultsense analysis; the interpretation, UK framing and recommendations are our own.