AI promises a huge transformation, but do not ignore the fallout

Artificial Intelligence (AI) is at the very forefront of investors minds. AI made a significant contribution to Microsoft’s record $62 billion earnings in the fourth quarter of 2023. This performance pushed the stock market value of the Microsoft above $3trillion for the first time. It is now approaching the size of the entire U.K. economy. Nvidia – arguably the most hyped stock in 25 years – is up over 350% over the last twelve months and is now the third largest company on the US stock market. Nvidia is estimated to produce four out of five of the chips needed to run advanced AI systems. AI is suddenly at the heart of every consultancy presentation. CEOs rarely make it through interviews without name-checking AI. Artificial Intelligence has economists excited at the prospect of an upturn in productivity growth as it transforms the knowledge and information management economy, and parks its tanks firmly on the lawn of a whole range of creative industries. It remains to be seen whether generative AI – widely seen as being the biggest digital breakthrough since the internet – has as many functional applications as claimed. But right now there are trillions of dollars being staked that it will.

Politicians – both past and present – are catching on. And not just because UK Prime Minister Rishi Sunak might be lining up his next job. He and the Chancellor, Jeremy Hunt – and their respective shadows on the Labour benches – are acutely aware that widespread use of AI in the delivery of public services is crucial if demographics and stubbornly low public sector productivity are not going to result in a remorseless higher rate of UK tax. This tax rate is already on track to reach an 80-year high. From the processing of citizen enquiries, smarter defence procurement, and healthcare diagnostics there is huge potential for AI to enable the UK state to do more with less. There will be use cases as yet unimagined. Former Prime Minister, Tony Blair and his one-time opposing number, William Hague last year released a report on the economic opportunities for the U.K. from being a global leader in AI. Whilst not underestimating the challenge such leadership will require, they are almost certainly on the right side of history in this judgement.

So far, so encouraging. And notwithstanding what I am about to say, AI is almost certainly a tool that will enable human activity to be more efficient – which in a resource-scarce world has to be good news. However this enthusiasm should not lead economies to sleepwalk into some of the challenging implications of AI adoption. There are much better qualified people than I to opine on the threats to stable politics and contented societies from deep fake imagery, viral digital content, and identity theft. All of which are being turbocharged by generative AI. However economists – sticking to their running lane – should also be mindful of the twin-sided impacts that AI will have on a services-heavy economy that currently creates 80% of U.K. wealth and jobs.

To illustrate this point one needs to see AI’s potential impact as analogous to the globalisation of the manufacturing sector in recent decades. This, and concurrent rise of China, had a huge impact on industrial sectors in Western economies. Globalisation removed the power of localised manufacturing monopolies, in turn removing wage-bargaining power from workers and communities. Rather than competing with geographically proximate companies in Cambridgeshire, a widget maker in Wisbech was going toe-to-toe with one in Wuhan. It was little surprise that salary differences – amongst other cost differentials – meant demand ebbed away from Western workers and moved East. The impact on community cohesion, waning support for mainstream economic policy, and the politics of Donald Trump and Brexit can all be traced back to this deindustrialisation of Western economies. Has AI the potential to unleash a second such wave of disruption? It must be a reasonable chance.

At this point mainstream economists will correctly note that overall welfare is unambiguously enhanced by trade. Trade brings greater choice for consumers, incentives to innovate and bares down on inflation. AI-enabled trade in services looks set to be a tantalising example of history rhyming, if not repeating. Trade in services is currently just 14% of global GDP, whilst trade in manufactured goods is more than 50%. The transition of trade-light sectors like professional services, creative services and knowledge management – to greater trade density – will unfurl similarly disruptive economic impacts.

Is this an excuse to introduce trade barriers either in the form of cross-border tariffs, differentiated professional accreditation, or divergent licensing standards? I suggest there will be intense pressure from workers in sectors impacted by AI to introduce just such barriers to trade. This economic debate is centuries old. AI looks poised to be the latest frontier. It is unlikely to be the technology that unleashes mass unemployment. Like its disruptive forebears it will change, destroy, and create jobs. What will be interesting is whether politicians bruised by the recent backlash to globalisation are dissuaded from throwing their economies headlong into the AI revolution.

The UK’s AI safety summit last year at Bletchley Park focused, sensibly, on guide rails, standards and regulation – all of which are necessary to foster this emergent technology. This is particularly the case with generative AI which takes existing creative material as its source material for new, repackaged content. We can already see the commercial world of media, gaming, art and literature positioning their legal teams to challenge this unfettered use of proprietary content. This is the latest module of a long-running dispute between content creators and technology firms. The line between protecting proprietary content and economic protectionism won’t be black or white. It will be multiple shades of grey. It is one of the reasons why advisory groups and lawyers are licking their lips at the enormous fees that will be thrown off from arguing this interface.

For modern economies, largely struggling with productivity growth for the last two decades, AI has enormous potential to break this trend and address many long-standing problems. But the inevitable losers from this transformation will not go quietly into the night. Governments need a plan to deal with the fallout.

Published on The Times: 26/02/2024