Barring an asteroid strike, Keir Starmer is going to be the UK prime minister in three days. Given the lead in polling, I’d probably bet on him over an asteroid, too.
Labour will come into government with a broken state, a flatlining economy and no money. A thin manifesto and enormous parliamentary majority means the party will almost certainly end up stretching further afield for ideas about how to deal with that trilemma from hell.
So let’s try and offer some.
Free our data
In March 2006, back when the Guardian technology section was a physical supplement in Thursday’s newspaper, we ran a campaign to “free our data”. We wrote about government-owned and approved agencies such as the Ordnance Survey, the UK Hydrographic Office and the Highways Agency collecting data on our behalf. We asked: “Why can’t we get at that data as easily as we can Google Maps?”
The campaign was, over the years that followed, a mixed success. Across the public sector, a new norm was created that government data should generally be made available to the public when possible. It almost certainly influenced the direction of the gov.uk project, putting open data at the heart of the state’s digital footprint, and a glance at the top-level data.gov.uk website shows how much work has been done to that end.
Someone born on the day of that campaign’s launch will be voting for the first time on Thursday. Yet some of the most valuable pieces of our digital infrastructure are still locked up, behind restrictive terms or expensive paywalls.
The Postcode Address File is one example. It holds 1.8m postcodes and almost 30m postal addresses, and is the ground truth for how we navigate the country. It was privatised along with Royal Mail, but remains tightly controlled by the state, with access charges regulated by Ofcom and a unique license for the public sector to use it at a flat cost.
Freeing our data is the right thing to do, but successive governments have viewed it as expensive: giving up a valuable revenue stream, in the name of abstract concepts. But a Labour party looking for growth and state renewal over the next few years should recognise that if a government dataset is valuable enough to be worth charging for, it’s even more valuable if it can be built on, improved and reused.
Similarly, much of the data that has been freed in the last two decades has been released under non-commercial licenses. The public is naturally squeamish about changes that could be seen to be “selling our data”, but offering state data for free to hobbyists and charging a license for commercial use is the worst of both worlds: the data is still sold, but the only businesses able to extract a commercial advantage from it are those already big enough to pay the fees.
There is more behind such a change than just the nebulous promise of economic growth; there’s also the simple fact that the government’s data is very good, and eliminating daily annoyances from people’s lives helps. I speak from experience: my flat was built in 2020, and for the first year I lived in it, I was functionally invisible to most e-commerce. Those who had paid through the nose to license the PAF could deliver to me; everyone else had to call and beg for directions, while they waited for changes to propagate through various inferior free databases that provide the same information.
Like Nixon going to China, a Labour government at the height of its popularity may be the only one that could sell such a change to the British people. Free our data, boost growth, and strip friction from our daily lives.
… Even the medical stuff
This is worth breaking out, because it’s far more controversial than simply offering access to Ordnance Survey maps. NHS England is one of the largest unitary health providers in the world. Its pursuit of clinical excellence is globally respected, even after 14 years of misrule; and its alliance with researchers across the country is foundational.
Pharmacology is on the cusp of a revolution. Vaccine development was pushed a decade ahead by the race to protect against Covid, and mRNA vaccines that can beat the next pandemic before it’s started are the next frontier. Gene therapy has gone from science fiction to dangerous experiment to fact of life, with rare genetic conditions now able to be cured, permanently, with a single course of personalised medicine. The cost of genetic sequencing continues to drop, allowing those same conditions to be diagnosed at birth rather than years later when the symptoms are damaging and permanent.
The NHS should be at the forefront of such research. For many types of genetic disorder, involving just a single base-pair mutation, a cure already exists in theory, but the practical realities of creating, testing and certifying it is beyond the ken of private industry.
Yet a disease that affects one in a million children pops up in the UK once every 18 months. With free genetic testing offered at birth, and a state capable of taking the long-term view of costs, a treatment could be offered to every single one of them.
It’s not only the right thing to do; it also helps the country at large. A £1m, or even £10m, project to treat a genetic condition in a single child is nothing compared to the cost of supporting someone to live their life with a profound disability.
And a cure doesn’t have to stay in the UK. A very different type of medical success, semaglutide, has been so successful it has reshaped the Danish economy. The drug, sold under names such as Ozempic and Wegovy, was developed by Novo Nordisk. In 2022 two-thirds of the country’s economic growth was due to the pharmaceutical industry, essentially down to its phenomenal popularity.
Get tough on Big Tech
after newsletter promotion
The groundwork has already been laid for the Digital Markets Unit, the arm of the Competition and Markets Authority (CMA) tasked with enforcing the digital markets, competition and consumers bill, a wide-ranging piece of legislation that serves as Britain’s equivalent of the EU’s Digital Markets Act. But the next secretary of state still has leeway to judge exactly how it gets enforced.
One of the first things on Labour’s plate will be sending the formal instructions to the CMA kickstarting that process. Doing it as a priority in the summer would be valuable, because the European experience shows that much of the response to competition enforcement consists of delaying actions; equally as important are the quirks of wording in those ministerial instructions that will set the tenor of the next few years of squabbling.
The operative word in all those weighty proper nouns above is “markets”. The focus of this legislation, in the UK and EU, has been on the aspects of big tech that skew the free market. There are businesses that cannot exist because the operation of the App Store, Amazon’s Marketplace or WhatsApp prevents them from doing so. Those companies’ arguments are invariably that the limits they impose are for their users’ benefits; the DMCC says: “We’ll be the judge of that.”
At worst, the changes it wreaks will be zero-sum transfers from large tech companies to smaller businesses. At best, they could help kickstart a new wave of entrepreneurship, at the same time as delivering a free upgrade to every smartphone in Britain.
Take AI government seriously
The state of the art of LLMs is already good. Better, I think, than is appreciated: as the awe has worn off, most of us have got very good at spotting the hallucinations, the stilted turns of phrase and the ease of jailbreaking the systems, and we have done so faster than we’ve worked out how to take the raw power of frontier AI models and use it to augment our own abilities.
By the end of Labour’s first term, it will be impossible to ignore how powerful this technology can be for the tasks of government – tasks which are, on an essential level, about coping with the vast quantities of information received by the state and working out how to efficiently manage and act on them.
The trick is to get started before then. Labour needs to know how, when and why it wants to use LLMs to boost the state so that it’s ready to go at the point it decides they’re competent enough to help.
For some functions, that moment may have passed. If the civil service is like any other knowledge employer, some proportion of its staff is already using ChatGPT or Claude to help proofread emails, pull together first drafts of memos or finesse lines to take; bringing that in-house would let the government ensure systems are being used safely and responsibly, while also expanding access to others who may be able to use them well but weren’t willing to stump up the cost.
In the future, more will open up, including the most dangerous part: interacting with the state. It will be a long time before any government could – or should – hand life-changing decision-making about individual citizens to an AI system; but you only have to fill in a few forms applying for disability living allowance, or an education and health care plan for a disabled child, to know that the current state is already a faceless machine with inscrutable motives. If an AI system can help people navigate that machine, rather than act as a further barrier to accessing help, then it could transform people’s relationship with the state.