Tony Blair says 40% of public sector work can be done by AI. Is he right?

Former prime minister Sir Tony Blair launched a policy paper this week suggesting that government departments could cut workforce time by up to 40% using artificial intelligence (AI).

The report from Blair’s think-tank, the Tony Blair Institute for Global Change, suggested that the UK public sector could save 20% of workforce time – but that the Department for Work and Pensions (DWP) specifically could save up to 40%.

But is the 40% figure even remotely plausible? Yahoo News spoke to experts in automating public sector organisations to find out.

The Tony Blair Institute for Global Change said that the adoption of AI across the public sector could save £10bn a year by the end of the current parliament. The think-tank’s white paper singled out the DWP because it is reliant on forms and paperwork, with fraud and error costing the welfare department £9bn a year.

The think-tank also suggests that harnessing AI to streamline and automate processes could cut backlogs. This would involve using algorithms to streamline application and assessment processes and improving prioritisation, and to cut down on repetitive tasks.

Meanwhile in job centres, an AI personal assistant can help claimants match to jobs, training and advice, and behind the scenes AI systems could help to spot benefit fraud (something which is already being done in the financial sector).

Blair said in an interview with the Today programme: "Government's all about process. You can automate a lot of these processes. In the DWP where we've done a deep dive with one of the UK's leading AI companies, we believe that you could automate 40% of the tasks there. And also, by the way, give a much better service to people."

So what do experts think?

The 40% figure is believable, says Bill Conner, CEO of Jitterbit, and a former adviser to security service GCHQ.

Conner, who helped move the UK to e-passports, says that the claim itself isn’t unbelievable, but it depends how AI is used.

“We have direct experience with our enterprise AI and we have seen savings of up to 40%," says Conner.

He adds that AI is particularly powerful in automating processes and integrating systems as large organisations on average have about 1,000 apps he says, but only 28% are integrated.

“Blair's claim is definitely not unbelievable when talking about AI capabilities. However, AI is not a one-size-fits-all solution and could even create more work than it saves if not implemented in the right way,” he observes.

Why the civil service poses particular problems

Dr Clare Walsh, director of education at the Institute of Analytics (IoA) and an AI expert who studied under world wide web inventor Dr Tim Berners-Lee, said that such broad-brush predictions should be taken with a pinch of salt.

“The idea that 40% of public sector work could, in theory, be done by AI is not dissimilar to the statistics cited over 10 years ago in an influential Oxford study, for the amount of work that would be performed by machines today. That prediction did not come to pass," she says.

Dr Walsh adds that AI holds the potential to automate tasks like scheduling appointments, but there are problems trusting confidential data to AI models.

"Nobody can submit government information to a machine that is housed abroad with no rights over the data once it’s left the government system and joined the millions of data points being used by a range of customers, including criminals," she adds

Dr Walsh says that small steps such as automatically taking notes, or helping to find information can add up to large time-savings, but that the Civil Service itself may pose unique challenges.

"Another barrier is that the civil service has some of the most complex chains of responsibility and accountability of any organisation, with systems that predate modern times. Of all the organisations in the country, the civil service would seem to be one of the most challenging to bring into the AI world."

Savings from updating old IT systems

Large savings are not implausible in the public sector, because many organisations are still struggling with very old IT systems, says Kam Patel, area VP, public sector at digital workflow organisation ServiceNow.

Government research in 2021 found that half of the money currently spent by the UK government on IT is on maintaining outdated legacy systems.

Patel says, "AI can drive improved efficiency by streamlining workflows and ensuring that civil servants can move their focus away from non-productive work." But there are risks, he adds.

“Leaders in the sector must always keep privacy and security in mind, being alert to the potential issues around AI. Citizens should always know when they are interacting with a chatbot, and technology leaders should monitor models for bias and other issues.”