British politics is hooked on flashy fake numbers – and the AI investment debacle proves it

## Summary
Keir Starmer launches the government’s action plan for AI at University College London East, January 2025. Photograph: Henry Nicholls/AP View image in fullscreen Keir Starmer launches the government’s action plan for AI at University College London East, January 2025. Anyone pitching a policy story knows the first question from an editor is: “What’s new?” The second is often: “What’s the number?” For high-profile government announcements or matters of current controversy, this is exacerbated by the fact that they are often covered, especially on the BBC, by political (rather than specialist) correspondents. When announcements fail to translate into reality, scepticism grows not only about the particular claim but about economic policy more generally, and about the potential for government action to change things.

## Article Content
Keir Starmer launches the government’s action plan for AI at University College London East, January 2025.
Photograph: Henry Nicholls/AP
View image in fullscreen
Keir Starmer launches the government’s action plan for AI at University College London East, January 2025.
Photograph: Henry Nicholls/AP
British politics is hooked on flashy fake numbers – and the AI investment debacle proves it
Jonathan Portes
A claim that the UK is attracting billions of pounds in AI investment has been debunked. That’s no surprise when our establishment runs on dubious ‘good news’
O
ne
trillion
dollars. That’s the amount of financial aid Gordon Brown triumphantly
announced
at the 2009 London G20 summit. (I contributed my own
two cents
here.) Except it
wasn’t exactly real
: the number was a mixture of already promised apples and aspirational future oranges.
So it should hardly be a surprise that when ministers proclaimed last year that the UK was attracting billions of pounds of new investment in AI, they were being more than a little economical with the truth. As a
Guardian investigation
revealed, much of it turns out not to be new at all: existing datacentres rented rather than built, a supercomputer site not yet even started, promised investments that might never arrive and claims of job creation that have little or no connection to reality. The headline numbers are impressive. The underlying reality rather less so.
Nor should this surprise anyone who thinks for a minute about how the economy actually works. Companies do not make billion-pound investment decisions because a minister wants a photograph beside a server rack. Policy matters – sometimes a great deal. But its effects are typically slow, uncertain and hard to attribute to any single government initiative.
Politics, however, runs on announcements. And announcements work best when accompanied by large numbers. So, when a company expands its UK operations, for whatever reason, it’s a victory for industrial policy. Existing plans and possible future investments are bundled together, given a headline figure and presented as proof that government policy is working.
The short-term incentive for governments to do this is twofold. The most obvious driver is the need to be seen to be doing or achieving something. The dominance of the No 10 grid, with its imperative to have a “good news” story every day, simply entrenches this.
But another important factor is the love-hate relationship of the UK policymaking establishment – ministers, civil servants and media commentators – with numbers. Many, perhaps most, aren’t particularly scientifically or economically literate. But that discomfort produces an odd result: when presented with a precise-sounding figure, they treat it with excessive confidence rather than scepticism. The assumptions behind it are rarely examined too closely – especially when the number is politically convenient.
We saw another example of this with Shabana Mahmood’s claim last week that if her
“earned settlement” proposals
are not implemented, “we will see a £10bn drain on our public finances and further strain on public services”. This is a mix of the
deliberately misleading
(the care workers she wants to expel are not a “cost”, but rather a fiscal benefit, for the next 20 years) and the completely wrong (her proposals would in any
case not save this money
).
Politicians are by no means the only ones to blame here. They are not alone in encouraging this dynamic. The media plays its part. Anyone pitching a policy story knows the first question from an editor is: “What’s new?” The second is often: “What’s the number?”
For high-profile government announcements or matters of current controversy, this is exacerbated by the fact that they are often covered, especially on the BBC, by political (rather than specialist) correspondents. Frankly, they simply don’t have the subject-matter expertise to understand, let alone unpick, the numbers they are served up in the press release. Rigorous investigative follow-up, as in the case of the AI announcements, is rare.
There are big costs to all this. First, it distorts public understanding of what policy can realistically achieve. Governments do influence the business environment – through infrastructure, regulation, immigration policy, education and research funding. Yet these influences operate over long time horizons and very rarely produce immediate outcomes, let alone ones that can be quantified with hard numbers.
Second, it undermines public trust. When announcements fail to translate into reality, scepticism grows not only about the particular claim but about economic policy more generally, and about the potential for government action to change things. That scepticism is already widespread in a country where productivity growth has been weak for more than a decade and living standards have stagnated.
Finally, it distracts ministers and civil servants from the real, much harder work of economic policymaking. Announci

---

## Expert Analysis

### Merits
- But another important factor is the love-hate relationship of the UK policymaking establishment – ministers, civil servants and media commentators – with numbers.
- This is a mix of the deliberately misleading (the care workers she wants to expel are not a “cost”, but rather a fiscal benefit, for the next 20 years) and the completely wrong (her proposals would in any case not save this money ).

### Areas for Consideration
- Anyone pitching a policy story knows the first question from an editor is: “What’s new?” The second is often: “What’s the number?” For high-profile government announcements or matters of current controversy, this is exacerbated by the fact that they are often covered, especially on the BBC, by political (rather than specialist) correspondents.

### Implications
- That’s the amount of financial aid Gordon Brown triumphantly announced at the 2009 London G20 summit. (I contributed my own two cents here.) Except it wasn’t exactly real : the number was a mixture of already promised apples and aspirational future oranges.
- So it should hardly be a surprise that when ministers proclaimed last year that the UK was attracting billions of pounds of new investment in AI, they were being more than a little economical with the truth.
- As a Guardian investigation revealed, much of it turns out not to be new at all: existing datacentres rented rather than built, a supercomputer site not yet even started, promised investments that might never arrive and claims of job creation that have little or no connection to reality.
- Nor should this surprise anyone who thinks for a minute about how the economy actually works.

### Expert Commentary
This article covers government, policy, announcements topics. Notable strengths include discussion of government. Areas of concern are also raised. Readability: Flesch-Kincaid grade 0.0. Word count: 927.