After the Bell: Davos, Day 2, AI, AI, AI

Memory can be a terrible thing. As they say (of course, I can’t remember who “they” are), memory is a science; forgetting is an art. I can remember going to Davos when blockchain was the rage. It wasn’t that the blockchain was going to change the world, but it was definitely on the brink of bursting into the public consciousness. Having doubts was equivalent to ignorance; only the ill-informed would be so bold as to deny the inevitability of the blockchain.

How many people today earn a living on the blockchain? My guess is not that many: the whole notion was enthusiasm-heavy but application-light. And it remained application-light; the only thing that changed was that it became enthusiasm-light too.

To say that artificial intelligence (AI) is one of the dominant conversation pieces at Davos this year would be an understatement. However, it is difficult to ascertain the balance between enthusiasm and fear.

A very good example was the blog posted on the first day of the meeting of the World Economic Forum by the managing director of the IMF, Kristalina Georgieva, which seemed to be laced with foreboding and enthusiasm. A recent study by the IMF showed that AI would “affect” almost 40% of jobs around the world. But, she added, it could also “jumpstart productivity, boost global growth and raise incomes around the world”.



So it’s a good thing then? Well … not so much. The findings, said Georgieva, were striking: almost 40% of global employment is “exposed” to AI. What do “exposed” and “affect” actually mean? That jobs will be lost? Presumably not. That jobs will be boosted? Sort of. It’s all a bit woolly.

The IMF’s study suggests that roughly half the exposed jobs may actually benefit from AI integration, enhancing productivity. For the other half, AI applications may execute key tasks currently performed by humans, which could lower labour demand, leading to lower wages and reduced hiring. In the most extreme cases, some of these jobs may disappear, she said.

She also makes the point that historically, automation and information technology have tended to affect routine tasks, but one of the things that sets AI apart will be its impact on high-skilled jobs. The IMF study suggests the biggest losers will not — as usual — be the developing world, but the developed world. Since its application in the developing world will be less widespread, disruption will be proportionately more acute in the developed world.

I discussed the issue in an interview with Old Mutual CEO Iain Williamson, who remarked that there is a difference between “traditional artificial intelligence” and generative AI, which is built on large language models. 

Just to take insurance as an example, traditional AI, or machine learning, has already found its way into large segments of the industry. Things like engines for underwriting rules and weather modelling use this form of AI.

Generative AI has yet to be applied to the industry, but it’s coming fast, most likely in the user experience, such as handling customer queries by email. This is a commercially viable use case today, Williamson said. 

“Large language models can allow the experience to be tailored on the fly so that you interact with the customer in the way that they want to be interacted with. I don’t think we’re quite there yet. But I don’t think we’re far away from it either.”

Two important international figures spoke at the forum on Tuesday:  Chinese Premier Li Qiang and EU President Ursula von der Leyen, and both discussed AI. 

Li described AI as a “double-edged sword”, which is a pretty common perception. Von der Leyen was positively gushing, saying Europe’s future competitiveness “depends on AI adoption in our daily businesses, and Europe must up its game and show the way to responsible use of AI. 

“There are nearly 200,000 software engineers in Europe with AI experience, and that is a greater concentration than in the United States and China, and our continent also has a huge — a huge — competitive edge when it comes to industrial data”. Okay, so more opportunity than threat then.

My impression, for what it is worth, is that it’s far too premature to start tossing numbers around about how many jobs will be “affected” or even how those jobs might change. It’s that curse of remembering the blockchain and the art of forgetting the promises made then. The difference is that, unlike the blockchain, AI is enthusiasm-heavy and application-heavy. And that is a heady combination. DM