Think at London Business School: fresh ideas and opinions from LBS faculty and other experts direct to your inbox
Computers are amazing: they can process vast amounts of data, at an astonishing speed, and they are getting smarter all the time. It is now decades since they showed their ability to do humdrum calculations such as ﬁnding a cube root. Now they can plan the quickest route to drive from A to B, avoiding traﬃc jams, and tell us our arrival time. They can predict pretty accurately how many cartons of milk a shop is likely to sell in two days’ time. They can recognise faces. They can steer a car through heavy traﬃc.
This exponential growth in computing power will not stop. Robots will become increasingly deft at performing tasks we currently see as the unique preserve of humans. They will become more and more skilled in interrogating patterns of behaviour by individuals and organisations and then suggesting better ways that tasks can be accomplished. Artiﬁcial intelligence (AI) will surround us even more than it does today. Within a couple of decades, the tech evangelists maintain, it will be able to replicate everything the human is capable of – from holding a genuinely stimulating conversation to choosing the right clothes for your day ahead.
Well, maybe. But let’s be clear about what AI can and cannot do. Its “intelligence” is essentially an ability to process and build upon what has gone before. Machines are literally “trained” by exposing them to huge bodies of data –text, pictures, codified speech—that allow them to spot patterns and make predictions. This can lead to seemingly-creative outcomes (such as the famous Go! victory by a computer that involved a strategic gambit no human had ever attempted), but it’s a form of creativity that is conﬁned within a narrow set of boundaries: it is about drawing an inference from past experience.
So how does all this relate to the future of work? It is already commonplace to use a rudimentary form of AI called Robotic Process Automation (RPA) for automating and speeding up many back-office activities, for example scouring thousands of documents to ﬁnd relevant precedents in putting together a legal case. Increasingly, such work and a myriad of other tasks will be taken over by AI, and the humans who used to do them will be redundant.
Those performing more creative, less mechanistic tasks at what we can loosely call the top end of the employment scale should escape this cull of their jobs. Those who work in areas such as providing care for the sick and elderly or serving in restaurants should also continue to see demand for their skills: empathy still counts for something. Hence, as the well-worn argument goes, the increasing application of AI will lead to a hollowing-out of the middle in the jobs market, while those at the top and bottom should see their roles change but endure.
But what does the increasing application of AI imply for companies? Certainly, for a company to survive, it will have no option but to adopt labour-saving, cost-cutting technologies. It will have to strive to match the operational eﬃciency of its competitors, who will all be doing the same thing. And of course, it will also need fewer employees. But all this does is no more than get a company onto the starting grid.
To win the race, it needs to make decisions about which customers to target and what new products or services might be devised to attract them. Here, AI’s limitations are revealed. Decisions such as these require intuition, imagination – and, crucially, an ability to pull together items of information from many diﬀerent sources. Lateral thinking involves far more than computing power, however vast. No computer ever dreamed up a cool new brand.
Compare venerable British high street institution John Lewis and Amazon, for example. John Lewis is a retailer. So is Amazon. John Lewis will use computing power for routine tasks such as invoicing and stock control: it tries to achieve operational eﬃciency. But given Amazon’s vast resources, John Lewis will never be able to do more than play catch-up in terms of delivering a given product at a lower cost. In terms of sheer eﬃciency, it cannot beat Amazon.
So, what should John Lewis do? Clearly it has to identify its unique qualities – qualities that involve the human touch – and concentrate on developing those, such as oﬀering advice and allowing a product to be inspected, tried out and compared, before it is bought, all within an agreeable environment.
The most important decisions that a ﬁrm makes will be about where to allocate its resources. And while AI can provide vast amounts of data about what has happened in the past, its predictive powers are limited and do not extend to making strategic decisions. For example, look at Facebook. Its algorithms feed its billions of users with material calculated to keep them engaged, providing huge audiences for advertisers looking to reach carefully segmented groups of individuals. But the company failed to spot the potential damage of users waking up to the reality that their personal details were being distributed far and wide; neither did it foresee that its product was being used as a medium for spreading untruths.
Facebook’s systems could not see the threats. It’s not simply that AI failed to spot the elephant in the room; AI was in a completely diﬀerent room. The threats were real enough, but it took the individuals at the head of the organisation to realise, belatedly, that these were issues that had to be tackled. Facebook’s shares have plummeted over the last year, and it is now recruiting thousands of people –with real human intelligence—to weed out the untruths and the fake accounts.
And that brings us back to the key point: deciding where to invest energy and resources requires lateral thinking, intuition and creativity – areas where humans trump machines. Managers within companies will have to devote an increasing amount of their time and energy to these right-brain activities. And they need to develop a more sophisticated understanding of what AI is capable of doing – and its limitations. This is emphatically not saying that a good manager will have to be a programmer. But she or he will need to have suﬃcient understanding of any AI system at least to be able to evaluate the information coming from it: to what extend can an answer be relied upon?
To use a somewhat trite example, if my SatNav tells me to use a particular route for a journey, I want to know whether it is taking into account traﬃc jams and roadworks. Similarly, if a computer programme tells me that a particular company’s stock is undervalued and therefore worth buying, I want to know on what basis it’s making that judgement.
And consider this. If I’m a fund manager and I have a piece of software that indicates when a stock is cheap or expensive, it’s inevitable that I won’t be the only fund manager using the software. Thousands of my competitors will be doing the same. If that’s the case, any potential proﬁt from following the software’s advice is likely to disappear in an instant. The only way to show an above-average return will be by being a contrarian and taking investment decisions that go against the AI grain. As Terry Pratchett said, “Real stupidity beats AI every time.”
An ability to evaluate the output of AI, creativity, imagination, drawing strands of inspiration from disparate sources and a willingness to challenge orthodoxy – these are the capabilities organisations need to develop in a world where AI becomes increasingly widespread. But no less important will be the manager’s eﬀorts to encourage colleagues to give expression to these quintessentially human talents.
That will mean creating a corporate environment in which radical thinking and experimentation are fostered and nourished: mavericks should be given freedom to come up with new and sometimes crazy ideas. Some experiments will fail, but that has to be seen as simply part of the cost of exploiting creativity. In a static state, AI may give sound guidance on where to allocate resources in the short term: its output is rational. But the really important decisions – about how much to devote to research and development or to training, and in which areas – demand very human attributes.
And, as more and more information becomes available to an ever-expanding cohort of individuals in a ﬁrm, the role of managers will have to evolve. For generations, managers’ status was bolstered by being the conduit through which information was disseminated, and by their exercise of control. No longer. The value that managers can add will increasingly come from using “softer” attributes to motivate and get the most from their employees.
These human qualities will increasingly be at a premium within an organisation. Take the case of a General Practitioner (GP). When you visit your doctor, she will have every detail of your medical history, and she will have access to AI-based technology which will allow more rapid and more accurate diagnosis of your condition. Does that mean she is becoming an increasingly unnecessary intermediary? Not at all. Access to all that data allows the doctor to make quick and well-informed judgements about your health prospects. And crucially, it frees up time to build a relationship with the patient. Don’t underestimate the importance of this. Evidence suggests that people who sustain a one-to-one relationship with an individual doctor over time are likely to live longer than people who see a diﬀerent GP each time they visit a surgery.
The march of AI – in medicine, in education, in public administration, in charities, in organisations of all types – will not stop. It presents threats, but it also brings countless opportunities. Humans and all their distinctive qualities will become ever more important in the quest for success.