The People Have Spoken. Someone Should Tell the Parties.
The UK local elections have delivered their verdict, and it is the kind of verdict that arrives written in flaming letters three hundred feet high, which is unfortunate, because the people most likely to receive it are politicians, who have spent entire careers developing a remarkable ability to squint at three-hundred-foot flaming letters and say "well, the picture is nuanced."
It is not nuanced. Labour lost seats at a pace that would have been embarrassing in a general election and is, in local elections, somewhere between catastrophic and geological. Reform UK, Nigel Farage's party, gained seats at an equivalent rate, picking up councils and councillors with the cheerful enthusiasm of someone at a jumble sale who has just discovered everything costs 50p.
To put this in perspective: Labour swept to power less than a year ago with one of the largest parliamentary majorities in British history. The party received this majority from a public that had grown exhausted by fourteen years of Conservative government and was prepared to vote for almost anyone who presented a vaguely coherent alternative. Labour presented a vaguely coherent alternative. It was a low bar. They cleared it.
What happened next is a masterclass in how to fritter away goodwill at speed. There is a particular political talent involved in persuading the electorate, within months of a landslide victory, that you are somehow worse than the people they just threw out. It takes commitment. It takes a specific kind of genius. Labour has it in abundance.
Reform UK, meanwhile, has achieved something remarkable: it has become the repository for everyone who is angry, which, given current circumstances, is quite a lot of people. This is not, it must be said, entirely to Reform's credit. Catching votes in a bucket during a downpour does not make you a skilled meteorologist. But here they are, bucket full, looking pleased with themselves, which is probably the correct emotional response.
What we are watching is the slow-motion implosion of the two-party system, which has dominated British politics since approximately the invention of trousers. The Conservatives have collapsed in one direction; Labour is collapsing in another; the Liberal Democrats are making steady, methodical progress that nobody seems to be able to get very excited about, which is probably their natural state; and Reform is hoovering up the dissatisfied middle ground with an enthusiasm that suggests they have not yet worked out what to do with it.
A Modest Proposal
Which brings me to the obvious solution, the one that nobody in politics will suggest because it is simultaneously too sensible and too threatening: what if we outsourced the validation of government decisions to an AI?
Not the decisions themselves, you understand. That would be going too far, and besides, AI systems currently have enough trouble agreeing on whether a hot dog is a sandwich without being handed the NHS budget. But the validation. The basic sanity check. The "does this policy actually serve the country, or does it primarily serve the interests of the people proposing it?" test that has, in practice, been largely optional for the last several decades.
Consider how such a system might work. A minister announces a policy. The policy is submitted to the Validation Layer, which cross-references it against publicly available data on outcomes, equity, long-term fiscal projections, and the observable fact that the minister's family owns shares in the relevant sector. The system returns a score. Policies that genuinely serve the public interest proceed. Policies that primarily serve the minister's golf club return a flag marked CONFLICT OF INTEREST: PLEASE TRY AGAIN.
This is not, to be clear, a proposal to replace democracy. It is a proposal to add a unit test suite to democracy. Democracy, as currently implemented, has a great many features and an almost total absence of automated testing. Bugs ship constantly. Some of them have been in production for years. A few of them have been in production since before living memory and are now considered features.
The objection will come, as it always does, that AI systems have biases, that they can be gamed, that whoever trains the model determines its values, and that this is therefore merely replacing one form of corruption with another form of corruption that runs faster and costs more to maintain. These are entirely fair points. They are also an exact description of the current system, which runs on human biases, is enthusiastically gamed, and is trained on a dataset of incentives that rewards self-interest at almost every turn.
At least an AI validation layer would be auditable. You could read the logs. You could, in principle, understand why a decision had been flagged. With the current system, you cannot even establish why certain decisions get made at all, because the reasons are distributed across a network of lunches, relationships, and ancient obligations that would take a team of forensic historians several years to untangle.
The UK local elections are, at bottom, a message from a public that has concluded the current system is not working for them. They are not wrong. The question of what to replace it with is harder, and the answer "Reform UK" may turn out to be approximately as satisfying as every previous answer that arrived wearing the clothes of the new while carrying most of the luggage of the old.
In the meantime, the flaming letters remain visible from space. The politicians are still squinting. And somewhere in a data centre, an AI is quietly waiting to be asked the right question, which is a state of affairs it shares, as it happens, with most of the electorate.