In 2017, Sebastian Thrun, an adjunct professor at the Stanford Artificial Intelligence Laboratory in California, unveiled a deep learning algorithm that can identify skin cancer as accurately as dermatologists with years of training. The algorithm was designed on a computer, but its creators want to make it available on a smartphone so people have screening for skin cancer right at their fingertips.
With over five million new cases of skin cancer diagnosed in the United States every year, it’s unlikely dermatologists there will be out of a job any time soon. However, it’s a stark reminder of just how capable technology has become. Its unstoppable progression doesn’t only threaten the traditional diagnostic skills of doctors. All the professions are going to feel the winds of change as machines take on more and more tasks.
Exactly what lies ahead for the professions is a major preoccupation for Prof Richard Susskind and his son, Daniel, a fellow in economics at Balliol College, Oxford. They are the co-authors of The Future of the Professions, which goes into some detail about the two potential scenarios facing the professions as a consequence of advances in technology.
"One is reassuringly familiar and essentially a more efficient and streamlined version of what we've got today," Daniel Susskind says. "For example, doctors using technology to speak to patients by Skype or architects using software to design more complex buildings.
Transformational
“The other is a very different proposition because it’s transformational and is effectively the active displacement of professionals by increasingly capable systems and machines. The Thrun algorithm at Stanford is an example.”
For now, Susskind predicts that these two futures will develop in parallel but eventually, he says, the second one will dominate as technology will offer new and better ways of solving the problems that professionals have traditionally solved in the past. Susskind describes this shift as “an existential challenge for the professions”.
He says many of the tasks that were once the exclusive preserve of doctors, lawyers and accountants are not going to remain so, mainly because they are “antiquated, opaque and no longer affordable”.
In an internet society, he argues, people neither need nor want professionals to work as they did in the last century. As a result, there will be a fundamental change in how specialist expertise is made available. “Our research suggests that people in search of professional services are actually looking for a reliable solution or outcome, rather than a trusted adviser per se,” he says.
Susskind is not suggesting that the end of work is nigh or that the professions will end up on the scrap heap. In fact, his advice to young professionals and those thinking of pointing their offspring towards the professions is to stick with it.
“There will always be a need for domain expertise so there is still value in taking the traditional academic or training route,” he says. “Very often now, it’s a combination of technical and domain expertise working together plotting the future. If you want to know a lot about medicine or accountancy or the law, the traditional route is still the best way to get it, but don’t imagine that’s it and from there your career will be a steady linear progression. You’ve got to be more open-minded and be prepared to think and move laterally in the world of work that’s coming. However, there’s not going to be a massive big bang, it’s going to be more incremental.”
Mass unemployment
So much for the youngsters. But what about older professionals, who still have a long way to go to their retirement?
In Susskind’s view they are not about to experience mass unemployment, but they are going to find themselves faced with redeployment and extensive retraining as the 2020s unfold and different skills are needed to do the new tasks that will emerge.
“Let’s take Thrun as an example,” he says. “What really interested me is that Thrun, a pioneer and expert in artificial intelligence with a wildly different set of skills to a traditional doctor, is able to design a system that can outperform leading medical professionals. That in a nutshell is the spirit of the observation we’re making – that the sort of people that are going to be involved in solving the problems traditionally solved by professionals are going to look very different in the future.”
Asked what skills will be useful in the years ahead Susskind identifies two possibilities.
The first is learning to do the things that machines and intelligent systems are not good at such as certain types of problem solving, creative processes and inter-personal tasks. The second is learning to design, build and operate the machines themselves.
Making these strategies work, however, requires educators and professional bodies to address skills gaps in these areas with some haste. Machines may struggle with practical tasks such as cutting hair, folding laundry and making beds but they are well suited to encroaching into high-skills areas such as medical diagnosis, assembling legal documents and scouring complex financial transactions for irregularities.
“In the medium term we’re looking at redeployment with a lot of change to come for the professions in the next 15 years,” Susskind says. “In the long run though, we find it hard to avoid the conclusion that there will be a steady decline in the need for traditional professional workers and this is a view that polarises people. In fact the divide largely corresponds with current views on AI with those on one side seeing it as a new dawn and those on the other dismissing it as hype.”
