Better skills will decide the winners in AI race

If generative AI is going to transform our lives in the way experts predict, what skills will be needed to thrive in this new world? Will it be “hard” technology skills or “soft” people and communications skills? The answer is not obvious, judging by a session I chaired at a recent conference.

“Engineering, mathematical, computer-driven skills: it’s all about computer science. That’s the area you’ve got to be specialist in,” declared Alasdair Haynes, founder of Aquis Exchange, a challenger stock exchange.

Er, not exactly, replied his fellow panellist Anna Anthony, UK financial services managing partner at EY. “The skills our people will need will be the things that computers can’t do … the relationship-building, the ability to have a conversation, all the face-to-face stuff,” she told the Financial Services Skills Commission conference.

In truth, the difference between them was not quite as stark as it sounded but it highlighted the huge uncertainty over the impact of AI.

Some experts reckon it will reduce the demand for core technology skills. Emad Mostaque, founder of the London-based start-up Stability AI, has predicted that there will be no human computer programmers in five years’ time.

Few would go that far, but generative AI technology dramatically increases the productivity of human coders. If it continues to improve, could that ease the current shortage of coders and other technology staff?

Not really, says Matt Candy, managing partner for generative AI at IBM Consulting. Every company, from carmakers to law firms, is becoming a software business, building software applications alongside their existing products and services. So companies will continue to need more technology specialists.

But at the same time generative AI will “democratise” programming, with non-technical employees able to turn their ideas into code. Softer skills such as creativity, problem-solving, critical thinking and collaboration “are going to be critically important”, he says.

Quite what balance of these skills will be needed by businesses is hard to predict but one thing seems pretty clear: there will be a big training challenge. Companies will not be able to rely on the education system or poaching from their competitors, they will have to find or develop the skills to use AI effectively within their existing staff.

If the enthusiasts are right, the potential rewards are huge. The good news is that more UK companies are early adopters of generative AI than their counterparts in the United States, China, France and Germany, according to a recent PwC survey. The bad news is that the record of UK companies on training is dismal.

Spending is half the EU average and has fallen by a fifth per head in real terms over the past decade. This has exacerbated the UK’s skills shortage, which is only going to worsen as the population ages.

Poor training may be one reason why UK companies’ productivity is low, despite spending on IT being higher per head than in France and Germany, according to figures from Gartner, a technology research firm.

The answer, according to Andy Haldane, former chief economist at the Bank of England, must involve more spending on training and more collective action by companies to raise skills in their sectors rather than a self-defeating strategy of poaching rivals’ staff.

Some industries are already well down that track. In finance, for example, the industry set up the Financial Services Skills Commission in 2020 to address its widening skills gap.

Some AI experts believe employees won’t need that much formal training in the use of the new AI tools as these will be so intuitive. They will be like apps on your phone that you can work out how to use by yourself. Maybe. But how many of us, particularly those of a certain age, make the most of the technology we already have in our pockets? Training will surely be one of the key factors that will decide the winners in the AI race.