Once upon a time, if you wanted to learn a particular skill, you had to ask another person to teach you, or you had to teach yourself, writes Laura Cox, Senior Staff Writer at D/SRUPTION. 

This might have involved going to a library, or signing up for an apprenticeship. We do still learn in these traditional ways, but in the digital age, access to information has become instant. 

People are more knowledgeable about more things than ever before. But, as argued by Tom Nichols in his book,‘The Death of Expertise’, this has reduced the need for people to actually learn. 

The internet alone has absorbed the requirement for vast amounts of expertise, and now artificially intelligent systems are adding to knowledge displacement. 

For your average person, this is a positive thing. But what does it mean for the countless numbers of highly trained, educated experts in a broad spectrum of industry sectors? 

What happens to experts when technology – in particular, AI – can eventually deliver the same breadth of knowledge?

The evolution of expertise

An expert is somebody with a high level of specialist knowledge in a specific area or areas. John Straw, D/SRUPTION co-founder, describes knowledge as an absolute. Knowledge itself is objective, he says, but the retransmission of that knowledge is subjective.

Lawyers, for example, create law on the basis of collective expertise, but these laws are questioned constantly. What AI can do is deliver an average of millions of lawyers and millions of legal cases.

“This creates a mathematical robustness, and as a result your subjective view almost translates into a fact,” says Straw.

The ability of a human to assimilate large amounts of information is extremely restricted. According to Straw, among those who will be most notably affected are healthcare professionals, lawyers, and accountants. 

For example, in the healthcare sector, radiologists go through a number of scans to decide whether or not a tumour is malignant or benign. 
Despite their extensive training, the rate at which they can view and analyse scans is limited.

“A computer could do as much in one day as a radiologist could do in their entire career,” he says. “The computer’s ability to become more objective over time becomes larger and larger the more data it is fed. Human experts are limited by their ability to process amounts of data. AI is virtually unlimited in this.”

Another example comes from the travel and leisure industry. A human customer services rep might be able to tell you that the all inclusive hotel you want to book has a gym, but not what equipment that gym has in it. Even when combined with generally useful star reviews on sites like TripAdvisor, this is still a minute snapshot of subjective opinions. 

What AI offers is the ability to see the average ratings of all reviews, and then narrow this down to certain features or services.

Professor AI

If experts can be described as retaining, relaying and using specialist knowledge, then intelligent machines, for better or worse, are experts in their own right. 

The ability to digest and analyse huge datasets has supercharged not only the quantity but the quality of the information available to cognitive systems. 

They can work out what information is useful and suggest what it might be used for, and how it may be used – that, in essence, is exactly what human experts do.

The training of these systems is important. Currently, AI is ‘narrow’ – it fulfils a preordained purpose and, being trained by humans  means it can exhibit bias. 

With a large enough sample base of data, it’s arguably possible to exclude this bias. This is the general premise behind ‘the wisdom of crowds’, an old idea later written about in 2004 by James Surowiecki in his book of the same name. 

According to the wisdom of crowds, the many are smarter than the few. If AI can combine the collective knowledge of millions of situations and examples, then it embodies the wisdom of the crowd. That is, of course, if it is fed the right data.

AI currently aids human experts by augmenting their existing knowledge. Once AI can teach itself, this dynamic will change. Being an expert becomes less important, with the role increasingly involving management and monitoring. However, expertise is not the same as experience. 

While an AI can access cold hard facts, it can’t have or relate to an experience in the same way as a human. This suggests that the people we regard as experts will change. Instead of knowing answers themselves, experts will be those who ask the right questions. Talking of which…

What comes next?

For Straw, the next milestone is context. Current narrow AI lacks contextual awareness, and this is why humans remain an important part of the transfer and reception of knowledge. 

Straw explains this fundamental difference using a joke by the American comedian Groucho Marx – ‘This morning I shot an elephant in my pyjamas. How he got into my pyjamas, I don’t know.’

“This is a really interesting example of the way that AI works, because you and I both know that it’s highly impractical for an elephant to get into pyjamas. Because we understand the context, we understand its relevance. We know what that joke means,” says Straw. 

“But if you looked at it from an entirely objective viewpoint, you would think that an elephant could have got into the comedian’s pyjamas because you don’t have context. An AI system will really make a breakthrough when it can understand that joke.”

Ultimately, Straw believes that the death of the expert is fast approaching. He predicts that within the next two to three years, expertise as we know it will start to disappear. 

However, a lack of contextual understanding means that there is still a place for traditional, human experts. Experience levels will play a key dividing role between the experts displaced by intelligent tech and those who work alongside it. The less experience you have, the more replaceable you are. Better get learning.

• D/SRUPTION magazine is Europe’s leading publication for business leaders looking to deepen their understanding of disruption and what is takes to scale innovation in the 21st century.

To keep up to date with all the latest developments on disruptive technology, sign up for the free D/SRUPTION newsletter at https://disruptionhub.com/newsletter/