Who really wins in the AI revolution?

The speed of change in the world of AI is so intense that even some algorithms might struggle to keep up with it. So it was worth pausing for an hour or so of reflection and debate at our latest think ahead discussion about “Who really wins in the AI revolution?”

Taking part were Nicos Savva, London Business School Professor of Management Science and Operations, Vinous Ali, Deputy Executive Director at the Startup Coalition, a UK network of over 4000 startups, and Gabriel Straub, a London Business School alumnus and visiting lecturer, and Chief Data Officer for Ocado Technology, the business that works with a range of partners in the retail sector. The discussion was moderated by journalist Tim Smith.

The panel brought together a valuable blend both of analytical thought and practical experience. As Nicos pointed out: “I’ve been working on data science and even AI before it was a cool topic.”

A poll of those watching live set the context for the discussion. As many as 56% felt that startups and innovators were best positioned to “win” from the AI revolution, while the gloomy view among the audience was that the ordinary employee or member of the public was far less likely to benefit.

But as the conversation developed it became clear that a more subtle (and less binary) state of affairs is unfolding.

Nicos explained that initial research into companies’ use of AI revealed a kind of levelling up effect, with productivity gains and improvements in quality being enjoyed by lower ranking employees. Experienced people were already quite good at their jobs and saw less immediate benefit.

“Then the results completely reversed,” he added. “The benefit accrued almost exclusively to the most experienced people…there’s something to be said about the value of experience…AI seems to be a complement to experience, and the most experienced people benefit the most.”

Vinous felt that most people are quite optimistic about what AI can achieve, particularly in certain specific areas. “So, for example, in the health sector and how it can augment and offer better services…I think the opportunities in the public’s minds actually outweigh the risks. And we just have to hold onto that trust and…start integrating it into our daily lives.”

“As many as 56% felt that startups and innovators were best positioned to “win” from the AI revolution, while the gloomy view among the audience was that the ordinary employee or member of the public was far less likely to benefit”

How are London Business School students looking at this developing world of AI? Nicos said that two decades ago “a third of our students would go to consulting, a third to banking and a third to the blue chips of the world.” While banking and consultancy are still popular options, the big tech firms are now a big draw as well. And budding entrepreneurs are looking at AI and “trying to bring AI solutions to different areas of life, from healthcare to retail to consumer goods, to education…”

Perhaps unexpectedly, Vinous suggested the public sector was well placed to benefit from the AI revolution.

“Government and public services have the potential to be faster adopters than others in this space,” she said, “partly because, with public services, you have a workforce that is used to being trained and being upskilled. One of the startups in our network works with GPs to transcribe the conversation that you’re having, and what that means is a GP who has 10 minutes to see you for an appointment can actually maintain eye contact with you throughout the conversation rather than typing notes that perhaps don’t fully capture the conversation, or don’t fully capture the problem, which leads to a misdiagnosis.”

Gabriel also had a positive perspective. “I think there’s certain areas where AI will enhance humans,” he said. “But there are also some jobs that are really rubbish, that humans don’t want to do, where AI might be able to replace it, or automation might replace them, which isn’t necessarily a negative thing from a human perspective…”

Nicos made another subtle but important point. “I think the right unit of analysis is not really the job, because a job consists of many tasks,” he said. “In doing any specific job, you need to do typically 10 to 20 tasks. And some of these tasks would definitely be replaced by technology.”

This is not bad news. “Even if, say, you’re doing a job in which 50% of the tasks can be done by technology and 50% remains the domain of humans, that doesn’t mean that we need 50% fewer people doing their work, necessarily. Why? Because that assumes that the total amount of work that gets done remains constant…Take the example of coding. In coding, the technology is able to do maybe 50% of the work of a coder, but not 100%. That’s not to say we’re going to need 50% fewer coders. Why? Because it’s more likely that we’ll end up writing twice as much code with the coders that we have.”

“In my view, the benefit of some of this stuff is that it forces us to be really good at strategy again”

Things are moving fast. Vinous said there was a lot of hype in the AI world, which means that some basic, hard-headed questions are needed. “What is the product that you’re building, what problem is it solving for, and actually is there a market out there for it? Who is going to actually use this? Who’s going to pay for it?” Gabriel Straub agreed. “In my view, the benefit of some of this stuff is that it forces us to be really good at strategy again.”

As for the future, Nicos warned that the sort of big transitions caused by new technology have typically led to major disruption and economic dislocation. “I would hope that we have learned the lessons of the previous transitions that were not so smooth,” he said. “We really need to start thinking seriously about investing in a safety net, around allowing people to redefine how they contribute to society productively in a world where their old jobs no longer exist.”

And our moderator Tim also warned us not to be naïve about the new technology, or too rosy about our fellow human beings. “I think we sometimes make the mistake of thinking AI is bad but humans aren’t,” he said.

Some humility, as well as imagination and optimism, is called for. Vinous said: “At the start of such a huge transition I don’t think any of us can truly have sight of what’s coming next.”

You can watch the entire discussion here