In August, retailer Marks & Spencer announced it was introducing AI chatbots to deal with customer calls to its 640 UK stores and 13 call centres. The system M&S has adopted uses a combination of Twilio voice recognition software and Dialogflow, an AI developed by Google. It won’t actually deal with a customer’s enquiry; rather, it’ll simply redirect them to the appropriate store or department (much as many existing, non-AI automated switchboards already do). But the news was nonetheless greeted with consternation by those who predict a ‘jobs apocalypse’ brought on by the widespread implementation of AI.
While artificial intelligence doubtless brings many advantages – ask any research scientist about the benefits of being able to analyse huge datasets thousands of times quicker than a human could – many people are worried that AI (or more accurately, automation using a combination of AI and robotics) could spell the end of their jobs. And depending on what job it is they’re doing, they may well have a point.
A lesson from the past
If we look at the history of technology, there are many examples of machines replacing human beings – the onset of the industrial revolution saw weaving as a cottage industry go into rapid decline. More recently, the introduction of robotics has seen millions of production line jobs in factories simply disappear.
Now, thanks to AI, anyone whose work largely involves following a routine set of procedures and making a routine set of decisions – from call centre workers and CCTV operators, to property lawyers and bank clerks, to court reporters and radiologists – may soon find that a computer can do it better, more quickly and cheaply than they can. Former chess world champion Garry Kasparov – who was himself famously bested by a computer way back in 1997 – has described this as “the cost of progress”, telling the Tableau Conference Europe that: “We used to lose manufacturing jobs to machines – the only difference now is the people who are at risk have college degrees and Twitter accounts.”
That’s the bad news – but there are reasons for optimism, too. Firstly, many experts suggest that artificial intelligence will ‘shred’ jobs rather than destroy them and, secondly, the number of jobs that new technologies create also has to be taken into account when considering their impact on employment.
What does ‘shredding’ jobs mean? Take the case of a radiologist: a computer may be able to examine CAT scans for anomalies that may indicate cancer, but it’s unlikely any hospital would allow a treatment programme to be instigated without human oversight somewhere along the way. Very few jobs in fact, are entirely a matter of routine – most require some degree of human input. So it may well be that AI replaces tasks, rather than jobs – freeing up workers from mundane, repetitive aspects of their job so they can focus on ‘human’ elements such as customer relations.
More importantly, there’s the issue of the jobs that AI will create. Again, we can look to history here. Many of those former weavers took jobs as machine operators in mills; once the internal combustion engine arrived, many an ex-coachmaker found themselves working in a car plant. AI is set to become a huge industry in its own right, employing countless programmers, computer scientists and engineers – and that’s not to mention new jobs we haven’t even thought of yet.
In fact, a recent large-scale study carried out by accountancy giant Price Waterhouse Coopers concluded that over the next 20 years, some seven million jobs would be lost to AI, but that around 7.2 million new jobs would be created as a result of it.
“In some sectors, particularly where jobs involve low-level repetitive tasks, the net effect of AI may be negative, with the potential for some jobs to become extinct,” says Dr George Panoutsos, Reader in Computational Intelligence at the University of Sheffield. “But in the broader manufacturing sector, I would expect to see significant growth in terms of demand for new skills, new processes, new legislation and standards, which all would generate jobs.
“In the short term,” he continues, “jobs that evolve around all aspects of creating, maintaining and using AI systems (from engineering and computing, to legislation and ethics) would see some growth. It’s challenging to predict what type of new jobs would emerge, however, one aspect could be the need to have expert users of AI systems, acting as the ‘interface’ between AI and the general population.”
“We used to lose jobs to machines. Now people at risk have college degrees”
Similarly, while Bank of England chief economist Andy Haldane has spoken ominously of the “technological unemployment” that AI might cause, he’s also spoken of the “skills revolution” that could counter those effects.
Experts all agree that artificial intelligence is going to impact upon the jobs market. Quite what form that impact will take remains to be seen, but there’s probably no reason to worry about global mass unemployment just yet.
Oh, and those M&S workers affected by the implementation of AI? They’ve all been found jobs elsewhere in the business.
How to get a job in AI
If AI’s going to create lots of new jobs, what do you need to do to get one?
According to LinkedIn’s 2017 Emerging Jobs Report, two of the fastest-growing jobs at the moment are ‘machine learning engineer’ and ‘data scientist’. That’s because, while AI might create all sorts of jobs in the future, right now most of the jobs involve creating the AI.
Many major universities including those of Leeds, Sheffield, Birmingham, Nottingham and Kent now offer specialist BSc courses in artificial intelligence (or AI and robotics), while many more run post-grad courses to MSc level. So for younger people, that would be the obvious place to start.
If returning to full-time education isn’t an option, many online learning providers also run AI courses that can usually be studied at your own pace, including Udacity, Nvidia and Coursera. And there’s one online provider – AI4ALL – catering specifically for groups, such as ethnic minorities and children from low-income backgrounds that are traditionally under-represented in the computing field.
And if coding really isn’t your thing? Then in the managerial sector, areas such as process re-engineering, business case development, data ethics and IP rights management are all going to be big business as AI booms.
Russell Deeks is a freelance writer with nearly 30 years’ journalism experience, working across the fields of music, technology and science – which, he says, cross over more often than you might think. Despite the drawback of holding a degree in English & American Literature, he has been a regular contributor to BBC Science Focus since 2006.