America’s anything-goes model is increasingly being ridiculed. As part of the current data-centre boom (or bust), ProEnergy has sold 21 repurposed jet engines to power data centres, often in urban areas. In practice, this means bolting old 747 engines onto concrete pads. Local nitrogen dioxide levels jumped 79%, and a 4× higher cancer risk. This is what a regulatory vacuum looks like.
Meanwhile, the EU AI Act (2024), the world’s first comprehensive AI law, is drawing criticism as a paper tiger: not enforced, heavy on bureaucracy, and poorly timed during an AI arms race. Many expected the UK to take a third path: a lighter, trade-compatible version of the EU structure, influenced by the surprisingly strong China Interim AI Measures (2023–2024) and India AI Governance Guidelines (2024), both of which are pragmatic and principles-based.
But the political reality is different.
Passing a full AI Act looks deeply unappealing for the UK government. AI is hyped, misunderstood and irresistible to backbenchers searching for cheap populist wins. Any bill would be amended to death.
At the same time, it’s becoming clear that the UK already has strong legislation that applies directly to many AI scenarios. Take HR, usually held up as the “hot potato” that needs new rules. In reality, the Equality Act 2010 and the Employment Rights Act 1996 have stood the test of time. Employers understand them. The public supports them. And they’re fully capable of handling AI-driven cases.
So what’s likely to happen?
The government will publish light guidance, probably from DSIT early next year. Existing laws will be mapped onto AI use cases. And industry regulators like Ofcom, the FCA and ICO will take a renewed interest in holding West Coast tech to account… ideally before more jet engines end up next to school playgrounds.