IT Essentials: Technosolutionism on steroids

The 'wow' needs to turn to 'how?'

Oh my, how politicians love AI! It really is the answer to all their prayers.

Tricky decisions can be taken off the balance sheet like so many bad debts, shunted into the never-never and left to machines of loving grace.

Fan boy politicos get to unleash their inner geek and rub knees with their Silicon Valley idols, swapping buzz phrases and bon mots that one suspects neither really understand.

Skills shortages? It's what AI was built for. Unemployment? Ditto. Budget deficits? An optimisation problem. Climate change - no need to worry about that anymore. Poverty? Schmoverty. Potholes on the A24? Just wait ‘til you see what Robofiller can do (and what a fab hi-viz photo opportunity too).

So it was with interest that we followed Jeremy Hunt's budget announcements for the NHS.

First up was fitting MRI scanners with AI capabilities. Fair enough. Without knowing anything of the detail, harnessing the acknowledged pattern matching powers of AI to a machine that produces a mountain of data sounds like a solid use case.

Then we had the news of that Hunt would fund a £3.4 billion programme to modernise NHS IT systems, in full. Great, until the line he undoubtedly needed to include to sell it to his colleagues: "But it helps unlock £35 billion of savings, 10 times that amount." Really?

Inevitably AI takes centre stage, by "slashing the 13 million hours lost by doctors and nurses every year to outdated IT systems" and "potentially cutting by half form filling by doctors."

At which point the "wow" should turn to "how?". How exactly is all this going to happen? What are the trade-offs? Projections like these are somehow allowable "because AI". AI is technosolutionism on steroids. But is it really any more than the new 'lick your forefinger and stick it in the air'?

As IT leaders constantly remind us, change is about people, process and technology, in that order. But AI conveniently allows politicians to ignore the first two, or to somehow imply that AI will make up for deficiencies in them.

We've been here before, of course. The National Programme for IT (NPfIT), the cost of which ballooned to £12 billion, was meant to connect GP practices and medical records centrally, made possible by the pace of technological advance. Rushed through against the wishes of most doctors and health experts (people), the programme was badly managed from the start, lacking clear direction, proper project management or an exit strategy (process). In the end the only winners were the big integrators.

A few days earlier we saw another example of AI solutionism: government trials using AI to reduce the size of the civil service. Described by deputy prime minister Oliver Dowden as "potentially a silver bullet" to solve problems across government, it was derided as "magical thinking" by Stochastic Parrots author Emily Bender, who pointed out that the term "silver bullet" is generally used as negation, as in: "There is no silver bullet that will solve this problem," which is undoubtedly more accurate.

Is the civil service too big? Quite possibly. Could the NHS be more efficient? Most definitely. Does AI have a role in increasing the productivity of public sector organisations? Absolutely. But the NHS and the civil service are highly complex organisations with a remit to serve the whole population rather than a small number of shareholders.

AI might indeed prove extremely useful in optimising and redesigning processes, and revolutionising our public services for the better.

But to quote a sage from a different age: you say you want a revolution, we'd all love to see the plans.