Political Tech
Brussels is quietly softening its landmark AI Act before the ink is dry, leaving the UK's go-it-alone regulatory strategy exposed.
NewsOnScale Staff
June 26, 2026
When the EU's AI Act was finalized in 2024, it was celebrated in Brussels as a historic achievement: the world's first comprehensive legal framework for artificial intelligence, a model that other democracies could follow or at least orient themselves around. Britain, freshly post-Brexit and eager to demonstrate regulatory independence, chose a different path — a principles-based, sector-by-sector approach that avoided the EU's prescriptive rules while still promising accountability. The bet was that flexibility would attract AI investment without sacrificing public trust.
Now the EU is softening its own rules, and that bet looks shakier than its architects would like to admit.
## What Brussels Is Actually Pulling Back On
The revisions under discussion in the EU are not cosmetic. Reports indicate that regulators are reconsidering compliance timelines, scaling back obligations on general-purpose AI models, and giving member states more discretion in enforcement — moves that critics say gut the Act's most ambitious provisions before they've had a chance to function. The driving force is familiar: competitive anxiety. European officials have watched American and Chinese AI development accelerate and concluded that their rules, as written, risk making the continent a less attractive place to build.
This is a significant moment, and not just for European policy wonks. When the entity that created the most detailed AI rulebook in existence starts loosening it under industry pressure, it sends a clear signal about who is winning the argument over how AI should be governed — and it isn't regulators.
## Britain's Exposed Flank
The UK government has argued, with some coherence, that its lighter-touch framework is more adaptive and innovation-friendly than the EU's top-down approach. The logic was partly economic and partly diplomatic: avoid Brussels-style compliance burdens while staying close enough to EU standards that British AI companies could still operate across the Channel without rebuilding their compliance stacks from scratch.
But that positioning depended on a stable EU framework to triangulate against. If Brussels is now moving toward something that looks more like the UK's flexible model anyway — not by choice, but under pressure — Britain loses its point of differentiation without gaining the regulatory clarity that a more structured approach would have provided. It has neither the EU's institutional heft to set global norms nor a distinctive enough framework to stand on its own as a model for others.
More practically, British companies and civil society organizations trying to plan around AI governance now face a moving target on two fronts simultaneously.
## The Accountability Gap This Creates
What gets lost in the debate about competitiveness and regulatory arbitrage is the original purpose of AI governance: protecting people from systems that can discriminate, deceive, surveil, and automate consequential decisions without meaningful oversight. Those risks have not diminished because Brussels is feeling anxious about its AI market share.
The loosening of EU rules does not make AI systems safer or more transparent. It makes the political economy of regulation more visible — and more troubling. When both the world's most ambitious regulatory framework and its flexible alternative are being shaped primarily by competitive pressure rather than evidence of harm reduction, the public interest is not being centered in this conversation.
## What Comes Next
The UK now faces a choice it has been trying to defer: commit to a more structured, rights-based AI governance framework with real enforcement mechanisms, or accept that its regulatory environment will be defined largely by what industry can negotiate in Brussels and Washington. Neither option is comfortable.
What is clear is that the window for proactive governance is narrowing. Every month that passes without durable frameworks is a month in which AI systems are deployed into healthcare, financial services, criminal justice, and public administration under rules that may never actually take effect. That is not a regulatory strategy. That is an absence of one dressed in policy language.
Independent oversight bodies, civil society, and journalists covering this space need to keep asking the uncomfortable question: when the rules change, who asked for the change, and who benefits from the version that emerged?