The Week

The Week 21 April 2023

James Sweetland
Senior Researcher

Recess is finally over and Westminster’s back to business as usual. With all the talk being local elections, it’s felt a little light on proper policy. But here’s a few items you might have missed.…

First, the IFS has kicked off its new Pensions Review — a multi-year project looking at what the UK should do to ensure people are financially prepared for retirement. The launch paper makes a lot of points we’re familiar with — the pressures of an ageing population on public finances (with annual state payments to pensioners to rise by £100 billion(!) by 2070), poor understanding about how to manage complex pension pots, and the usual concerns that people just aren’t saving enough.

Perhaps most interesting is the data around self-employed savers. The IFS reports that the percentage of self-employed people putting money into a workplace pension has dropped from 46% in 1998 to just 18% as of 2020. This is all the more concerning given the rise in self-employed workers in the past two decades: from 3.3 million in January 2000 to 4.2 million by March this year. And even those self-employed people who do make pension contributions aren’t saving enough — the IFS cites research showing that 25% of those who save for retirement contribute the same amount in cash as they did 9 years ago: a serious drop in real terms. This is a worrying trajectory. We await the first main report from the Pensions Review in Autumn 2023.

We've also had a series of different speeches from ministers this week — all of which flagged up some potential impacts of AI. Economic Secretary to the Treasury Andrew Griffith talked about how it can identify fraud and market risks in financial services. Security Minister Tom Tugendhat spoke about the not-at-all terrifying potential of artificial general intelligence (AGI). And the Chancellor dismissed concerns about AI taking jobs — and called for a “super smart” approach to regulation.

Ministers have clearly grasped the opportunity promised by AI. Questions will nevertheless persist about how the Government views its role in the extraordinarily difficult task of managing the impact and risks of AI, given this technology’s evident potential to disrupt every sector and public service. The recent AI White Paper was deliberately light-touch, focusing on existing regulators using their “domain-specific expertise” to manage risks, with some central coordination. This has advantages — flexibility, a tailored response by sector, space for innovation — but risks missing cross-cutting problems that might emerge and which require a system-wide response. This approach would also seem to place the onus on problematic uses of AI being squashed once they’ve emerged (by regulators), rather than prevented in the first place (through more prescriptive legislation).

And onto this week’s (health-themed) reads…

First, to continue the tech talk, there’s some AI bubble-bursting in this article about ChatGPT’s potential in medicine. It’s the reflection of a US emergency doctor and VP of a health tech firm who tested the tool to try and diagnose patients. In some ways, it worked — at least “when I fed it perfect information and the patient had a classic presentation.” But its ability to understand more complex cases and spot the anomaly from the obvious ‘right decision’ was poorer. As he writes, when it comes to healthcare, “most actual patient cases are not classic.” Large language models won’t be replacing any public service specialists, it seems, but the potential to boost their productivity cannot be dismissed.

Secondly and moving on from the tech stuff, we also enjoyed this behind-the-scenes account of Alan Milburn’s time as New Labour’s Health Secretary between 1999 – 2003 in The Independent. There’s a whole bunch of fascinating insights here, but the stuff about pushing through reform really stood out. He describes trying to introduce more autonomy for patients in the health service and being told by a senior civil servant: "Minister, patients don’t want choice". When a pilot project let patients choose where they had eye operations, 75% picked a different hospital. Leaving aside the specific question of choice, there’s a lesson here about assumptions among those creating policy. Plenty to chew on for Wes Streeting, off the back of his speech today.