My IEA article on AI and government

AI offers great opportunity to the public sector. Ministers tell us they are spending on it to boost productivity and drive change. We were told of a £3.25bn Transformation Fund for public services for 2025-8, with £2bn being spent this year and £ 1bn next. The recovery of lost productivity was always planned to take time, with AI mentioned as a way of helping do it. We  were treated to another 5 year hockey stick, with extra spending only leading to substantial savings in the final year.  I never understood why we needed new investment in AI to recover 2019 levels of productivity when that was a pre AI era. The government should urgently recover the lost productivity on old technology and  consider what additional gains there can be by spending on new technology.

The government needs to live up to its words on AI and chart a way of using it that is fair to taxpayers and its staff. Throwing a few hundred million more into the IT budget and hoping the  suppliers will spend it well for us is not a good strategy. What is their aim? Which parts of government will benefit most? Is there a central strategy or do different parts of the public sector make their own way through the AI spending grounds? Much of public administration is amenable to more computing assistance, but it needs to be done in a way which keeps the good staff, allows them to direct and improves the lot of the users.

Now we can see the breakdown of the planned Transformation spending for 2026-7 we see £547 m is going on SEND, £ 271 m on Social Care and £150 m on asylum. This implies  half the Transformation money is going to three hotspots where budgets are under severe pressure and will not primarily be on AI. There is only £25 m this year for  civil service improvement, £210 m for HMRC computer schemes and £194 m specifically earmarked for various cross cutting digital programmes.

I have written before that efficiency and quality are two sides of the same coin. Higher quality and productivity have to be driven daily by senior managers. In government Ministers need to stress their importance. In order to achieve them the top managers need to learn daily from experience. What has gone wrong? What can we do better? What do we not need to do? Which complications mean more likelihood of mistakes and cover ups , even where their ostensible purpose is greater safety or accountability? Intelligent use of AI can clearly change the frontier between what you can let a machine do and what has to be done by a person. AI cannot substitute for hands on direction and responsibility for the service held by people.

When I was leading industrial companies I gave myself two important tasks as part of the job to highlight the importance to my teams of quality and efficiency. I asked to be informed of all serious complaints and accidents so I could see what was going wrong and insist on remedies, with a view to designing out the risks. I was also the questioner for the future, asking my managers all the time what we could be producing next that would be better, how we could improve service quality and customer relations, identifying inadequacies and antiquities in our factories and products.

Managers were often shocked to be told that most of what they are making today they will not be making in five years time. People get caught up in the here and now, but some have to look to the future. What look like good secure jobs today, backed by a steady flow of orders, can vanish tomorrow if leaders do no think ahead about what consumers will want and expect next. Making records in an age of downloads or selling carts in an age of cars do not create many jobs or need large factories.

Cost down or cost control is part of daily life of any company that wants to survive in a competitive market. Left wing critics think management is driven by an unhealthy desire to maximise profits, always looking for ways to do things more effectively. They need to understand that the prime reason for costs down is to provide customers with keener prices. This is followed by the need to control costs to stay in the market, protecting the company’s jobs. Put up prices too much and allow costs to be out of control like government and the jobs will be lost to competitors.

Profits too do matter, because out of the profits you can afford to research the future, launch the new products and services and above all pay for the new factories and computers you will need to deliver a good answer. Most profits do not go to rich shareholders. If no profits go to rich shareholders there will be no rich shareholders creating the company and the jobs in the first place.  Profits  go to new investment, to paying interest on borrowed capital, and to paying dividends to the pension funds and insurance schemes of the many as well.

AI is seen by some as a major threat to jobs, as AI will be able to do more things better and more cheaply than employing a person to do them. For any individual company there is no choice if they want to survive. If the computer is cheaper and better the company will need to use it more to stay in the market. Used properly society gets richer, meaning we will have more to spend on new products and services still not yet dreamed up. The cost savings from AI get recycled, helping growth and prosperity. Real wages go up as jobs change and total output and income rises.

Just look at the energy of Amazon, one of the world’s largest companies, that understands all this. Last quarter it reeled off an incredible list of all the things it was doing to widen its reach and serve its customers better. In the retail space alone it launched the faster supply of medicines to people with bigger discounts on price, expanded its ultra fast delivery service offering 30 minute same day delivery in some big cities, added millions of new products and many brands to its range of products for sale, launched Amazon Supply Chain services so other companies could use the advantages of its efficient supply and delivery systems, announced Proteus, an enhanced robot lifting loads in warehouses , and developed Alexa as a Shopping Assistant.

What did the UK government develop and launch to give users of its services faster and cheaper services? Why did it go on ignoring complaints and putting up with poor service? Why did it greatly increase the costs without improving the quality? Why does it always boast about how much it spends on a service but not about how much service it delivers? Who was the voice in government for the future, seeking to apply the new technologies to delivering better solutions to users that they wanted? Why do so many in government see such digitalisation as there is as a new way to enforce compliance and make things easier for the service supplier, not the user?

When I was a Minister I was shocked at the unwillingness of parts of the public sector to learn from mistakes and to improve what was clearly poor performance. I had to keep repeating the need to answer incoming letters promptly and well. Today emails go unanswered for months in some cases, produce lazy or unhelpful replies sometimes and sometimes go unanswered altogether. When a disaster occurs like the Horizon scandal, the contaminated blood scandal, the avoidable deaths in hospitals scandals, the dumping of refuse in the countryside and the rest there are attempts at cover up, arguments over responsibility, and long delays in providing remedies and compensation. If one of my companies had a worrying complaint I got them to settle urgently, to apologise, and to show the harm was not meant and the  remedy was good.