Data mining exposes the tension between EU alignment and AI ambition
When the EU sets rules that shape markets, supply chains and legal risk, the UK faces a choice: align, diverge, or drift. Too often, we drift. Text and data mining provides a neat example of how this drift might undermine our ambitions in AI, says Anand Menon
For some people at least, Brexit was about regulatory agility, the ability to move fast and steal a comparative advantage over the ponderous, risk-obsessed Brussels behemoth. Rachel Reeves hinted at it in her Mais lecture yesterday. Alignment would be the rule, but in areas with unique characteristics or strategic importance, autonomy might be necessary. AI is one such sector, singled out by the Chancellor as one of enormous economic potential
Yet economic potential does not imply regulatory nimbleness. Indeed, in some of the very areas where that freedom was supposed to deliver a growth dividend, we have ended up moving more slowly, with less clarity, and with fewer results than the EU.
Remember the pitch. Leaving the EU would let the UK “unshackle” itself from Brussels’ rulebook. We would be nimble, experimental, pro-innovation — the sort of place where fast-moving sectors like tech and AI could thrive. Regulation would be lighter, smarter, more tailored to British strengths. In short: sovereignty would generate prosperity.
In practice, however, “freedom” has too often meant friction. Whitehall has spent years managing the consequences of divergence (or the threat of it), while ministers have repeatedly discovered that big, trade-exposed sectors cannot simply pick up and ignore the EU’s gravitational pull. When the EU sets rules that shape markets, supply chains and legal risk, the UK faces a choice: align, diverge, or drift. Too often, we drift.
Text and data mining (TDM) provides a neat example of how this drift might undermine our ambitions in AI. In plain English, TDM is the automated reading and analysis of large volumes of text and data to find patterns — a building block for everything from fraud detection to medical research, and vital to the UK’s industrial and services sectors.
Economy-wide questions
This is not a niche copyright spat between tech companies and publishers. It is, increasingly, an economy-wide question about whether UK firms can use data confidently to build and adopt AI tools.
Here’s the awkward bit for the Brexit story: the EU has already made a decision. Across the EU, companies are generally allowed to use text and data mining for commercial purposes unless the copyright owner has clearly said “no”. That “no” can be signalled in a standard, machine-readable way — essentially a digital opt-out.
Whatever you think of that model — and reasonable people might pick holes in it — it has two virtues: it is legible, and it is settled. The rules exist. Businesses can plan around them. Rights holders have a clear route to say no, and a clear basis on which to negotiate.
The UK, by contrast, is still arguing about first principles. Our current statutory exception is narrow: it allows copying for text and data analysis only for non-commercial research, subject to lawful access. Everything else sits in a grey zone of licensing, legal uncertainty, or avoidance.
We have been here before. In 2022, government floated a broad UK exception that would have allowed TDM for any purpose — and, crucially, without opt-outs. It collapsed after a backlash from the creative industries. Since then, we have had consultation, debate, more consultation — and now, yet more delay.
Ministers have signalled that the politically contentious decisions are being pushed down the road.
Step back and think about what that means. The EU — caricatured for years as slow and bureaucratic — has an operational commercial regime with an opt-out. The UK — supposedly liberated to sprint — is still stuck at the starting line,.
And delay has distributional effects. Uncertainty tends to favour the biggest players: those with legal teams, lobbying capacity, and the ability to cut bespoke deals. Smaller businesses — the ones ministers say they want to champion — are left with risk and ambiguity. Creators, meanwhile, face a market where enforcement is hard, standards are inconsistent, and negotiations are fragmented.
Our own research for Microsoft underlines that TDM is already embedded across the economy — and that future AI adoption will depend on clearer rules and predictable access to data. It also shows the potential scale of the prize, up to £510bn of annual growth by 2035, and the potential cost of getting stuck in a war of words between two of Britains great industries (The point is not that one side should “win”; it is that the UK needs a framework that actually works in practice.)
Brexit did give the UK more room to move. But room to move is not the same as getting things done.
If the government is serious about growth in tech and AI, it needs to show it can translate slogans into decisions — quickly, clearly, and in a way that enables economic growth.
Otherwise, “take back control” will end up meaning something rather different: watching others set the rules while we debate them.
Anand Menon is a director at Public First and director, UK in a Changing Europe