The Future of Professions
Richard and Daniel Susskind predict the decline of today’s professions and discuss the systems that will replace them
In an Internet society, Richard and Daniel Susskind argue that we will neither need nor want doctors, teachers, accountants, architects, the clergy, consultants, lawyers, and many others, to work as they did in the 20th century. Their book, The Future of the Professions, explains how ‘increasingly capable systems’ — from telepresence to artificial intelligence — will bring fundamental change in the way that the ‘practical expertise’ of specialists is made available in society.
In an era when machines can out-perform human beings at most tasks, what are the prospects for employment, who should own and control online expertise, and what tasks should be reserved exclusively for people? The father and son writing team came to Second Home to discuss further.
Richard Susskind: We thought we should start off with just a little bit of history — how did father and son come to be writing a book together? For 35 years I’ve been working away and trying to drag the legal profession from the 19th Century into the 21st Century — arguing that technology must play a role in improving the way we gain access to justice, the way our courts work and the way lawyers practise.
I give talks regularly, and what kept on happening was that at the end of talks, a stray architect or perhaps a doctor would come up to me and say, ‘Richard, it’s interesting to hear what you’re saying, it’s happening in law as a result of technology, but we’re seeing exactly the same in our profession, too’.
I’d been working also a little bit with accountants and tax advisers, and it seemed to me after about 30 years that the thesis I was evolving about the way in which the legal profession would change, might have a far wider application.
At the same time, Dan was working at number 10 and he was looking at various issues to do with technology, and health policy, and justice policy — there was so much coming across their desks about the way in which the world was changing. Daniel and I started discussing what we were seeing and our feeling was there’s something very major afoot that technologies were not just working at the periphery of the professions, but might be fundamentally changing the professions.
So about five years ago we joined forces. Daniel and I had a great time working together, but throughout he felt that I wasn’t really pulling my weight. So if you can imagine my wife, his mother, receiving calls from Daniel saying, ‘Dad’s not working hard enough on this’, it’s a massive role reversal.
But we worked together for five years looking at different aspects of the professions. Some of this actually involved going out to meet thought leaders and market leaders who were changing the way that professional services were delivered, and a lot of this also involved us looking at written sources, books, on the web, and so forth.
The very, very broad thesis is that there’s two futures for the professions. The first future, we say, is reassuringly familiar — it’s when technology essentially streamlines and optimises the traditional way of working. So a doctor might see a patient over Skype, an architect, rather than handcrafting a plan or design, will use computer-assisted drafting software, or you might have teachers using online sources or whiteboards and so forth. All of these technologies we’re familiar with, entirely everyday technologies, but what they’re doing is streamlining and optimising the traditional way of working, not changing the way the professionals work.
We speak to most professionals that are quite comfortable with this that say, ‘Yes, this is the future of technology, it’s making me more productive, it’s making me more efficient’, and we saw that in our studies and our research. But we also saw in parallel a second future, which was fundamentally different.
The second future is that we find, through technology, entirely new ways of sorting out the problems for which our professions used to be the only solution. This is the most challenging aspect of our book, because if you take my home area of law, most people assume if they’ve got a legal problem, you go to a lawyer. We’re arguing [that] actually, for example through online servers, online corps, there might be entirely new ways of addressing legal issues.
If you look at medical diagnostics online rather than the idea of seeing a doctor through Skype, it’s a new way of delivering the professional service. It’s a way that often is threatening to the professional. That’s one of our running themes — very often the questions we’re asked are from professional people who want to challenge the idea that their occupations are under threat. We appreciate the threats, but we actually have another perspective — we’re more interested in the recipients of service, the patients, the students, the clients.
Our hope and our vision is that the second future is one in which people have far readier, easier, less expensive access to solutions to the problems for which our professions have traditionally existed.
“Our hope and our vision is that the second future is one in which people have far readier, easier, less expensive access to solutions to the problems for which our professions have traditionally existed.”
In preparation for the book, Daniel in particular, in the United States, did a bit of a world tour and spoke to leaders.
Daniel Susskind: We look at lots of case studies in the book, but to give you a flavour… In the legal world, last year on eBay, 60 million disputes arrived that were resolved online without any lawyers at all by using what’s called and e-mediation platform. Bear in mind that 60 million disputes — 40 times the number of civil claims that are filed in the entire English and Welsh justice system — were resolved without lawyers on this one website.
Again, in the legal world, the best-known legal brand in the US isn’t a traditional law firm, it’s LegalZoom.com — online document-drafting and legal advice platform. In education, Khan Academy, online collection of practice problems and instructional videos — I use it to teach my students maths and economics — it has 10 million unique users a month, which, in a way, is a higher effective attendance than to all the schools in England.
In the medical world the US Food and Drug Agency has said that by 2018 there’ll be 1.5 billion people with at least one medical app on their smartphone. Again, WebMD — online network of help websites, extensive guidance on treatments and symptoms — has 190 million unique visitors a month, which is more than to all the doctors working the United States.
“The US Food and Drug Agency has said that by 2018 there’ll be 1.5 billion people with at least one medical app on their smartphone. WebMD has 190 million unique visitors a month, which is more than to all the doctors working the United States.”
IBM’s Watson is a supercomputer that is owned by IBM, and in 2011 it went on Jeopardy — a US quiz show — [which is] about the same difficulty as Mastermind in the UK, and beat the two best ever living human Jeopardy champions. This was a system that, in effect, could answer a question on anything in the world, better than the two leading human experts.
IBM have been using it to help with the diagnosis of various illnesses and recommend treatments, especially in cancer and post-traumatic stress disorder. The most interesting thing that I’ve seen recently though is that it’s teamed up with the Baylor College of Medicine and is actively helping in medical research.
There’s a protein in our bodies called p53 and it’s called the ‘cancer killer’. There’s a huge research agenda looking to try and uncover the enzymes that can turn on this cancer killer. If you can find the enzymes, you can turn this protein on and it stops or slows down the growth of cancer.
Until IBM Watson came along, in the medical literature there were 33 possible enzymes that the medical profession knew about. The problem was that the literature was 70,000 publications big. The estimate was that it would take a good medical researcher about 34 years to read this entire literature — it was really stagnating. IBM Watson scanned the entire medical literature, those 70,000 and apparently six million others which were kind of tangentially related to that area, and was able to generate six entirely new enzymes that could potentially turn on p53. This is a really interesting move from just doing diagnosis to doing medical research.
“IBM Watson scanned the entire medical literature, those 70,000 and apparently six million others which were kind of tangentially related to that area, and was able to generate six entirely new enzymes that could potentially turn on p53 – the cancer killing protein.”
That’s what we see across the professions — these systems and machines not only making things more efficient or effective, but doing things in fundamentally different ways.
Associated Press, in 2014, started to use algorithms to computerise the production of earnings reports. It was able to produce 15 times as many earnings reports using these systems than it was when it relied upon traditional print journalists — a fundamentally different way of doing things. That’s not streamlining and optimising the work of a financial journalist, it’s doing it in a completely different way.
So that’s what we see — we see these two futures at the moment developing in parallel. This future of making things more efficient, but there’s also this future of doing things in very different ways.
Richard Susskind: In 1996 I wrote a book called The Future of Law. One of the things I was going on about was email. I said that the dominant way that clients and lawyers would come to communicate in the future would be via email — that was one of my major assertions. The Law Society of England and Wales at the time said I shouldn’t be allowed to speak in public. They said I was bringing the legal profession into disrepute by suggesting that lawyers and clients would use email.
What we suggest in the book is that reaction, although farcical in retrospect, is the very same reaction we get to our discussions of new emerging technology. We also look at big data, the way in which very large bodies of data, say, in medicine, can actually yield predictions, diagnoses, patterns, trends, more powerfully often than medical science.
We look at affective computing, the idea of machines that can both detect and express human emotions. There are now systems more accurate than any human being can look at a human face and see a smile, and tell whether or not that smile is a fake or a genuine smile.
Machines, by 2020, will know what kind of mood you’re in — your jacket will give you a little hug when there’s a ‘like’ on Facebook and so forth. [Things] will be connected up, not only to the Internet of Things with chips embedded in everyday objects, we’ll have chips embedded in ourselves in the 2020s.
So we’re moving into a very different world, wherever we looked, in all the professions. And it’s early days, we’re seeing not simply this optimisation of the old ways of working, we’re seeing transformation.
We started off thinking we were writing a book about the future of the professions, but actually we ended up writing a book about how it is in the future we will produce and share expertise in society. Because that’s what the professions are all about — whether you’re a doctor, a lawyer, a teacher, an architect, a consultant, a clergy, a journalist — what professionals tend to have is knowledge, experience, skills, expertise, and those that consult them, don’t.
The only way as non-experts that we’ve managed to solve problems in medicine, law, architecture or everyday challenges, is by going to the gatekeepers, the professionals. What we’re seeing now is — and we identify — six alternative models for producing and sharing expertise in society. That’s why, understandably, many professionals feel threatened by the thesis — it’s also why we feel excited by the thesis, because there’s genuinely going to be more access.
Daniel Susskind: The traditional professions at the moment are creaking. Most people in society simply don’t have affordable access to the expertise of first-rate professionals or indeed any professionals. Most people simply don’t have access to legal advice, they don’t have access to good healthcare, they don’t have access to good education.
The professions are under-performing, they’re opaque. Lots of the things that the professions do is very, very difficult to understand, and sometimes that’s because what they do is genuinely too complicated for people to understand, but other times — if you look at the way lawyers dress, the language doctors use — there’s intentional obfuscation there as well.
“Each profession that you look at simply use old-fashioned ways of producing and sharing information and knowledge, despite the existence of feasible alternatives. Most courts, for example, in Britain still have fax machines.”
And they’re antiquated. Each profession that you look at simply use old-fashioned ways of producing and sharing information and knowledge, despite the existence of feasible alternatives. Most courts, for example, in Britain still have fax machines. In the book we talk about the print-based society, a society where if you want access to information, knowledge, skills, expertise, you go to the professions. As we move from the print-based society to an internet society where there are entirely new ways of producing and distributing expertise in ways that don’t rely upon the traditional professions, the fact that the traditional professions are doing such a bad job is a real reason to embrace these alternatives.
Richard Susskind: We can imagine a world where what we call practical expertise — the knowledge, the experience, the know-how, the skills of experts — can actually be made available online. Imagine that world, and that knowledge could be made available online as a chargeable service so you could actually sell as many lawyers and others already doing their electronic versions of their service.
It could be made available by governments and other charitable bodies who control the content, but make it available at no cost, or it could be made available in the principle of a commons — in the spirit of open source software, in the spirit of Wikipedia — as a body of practical expertise to which people could contribute and draw from and edit, owned and controlled not by major corporations, professional organisations, charities or governments, owned and controlled by the users themselves.
Of course the other option is we keep the professions and the professions continue to be the gatekeepers and if you need access to insight, guidance, help and advice, you have to go and pay intermediaries. We ask the question, ‘If you know nothing about your own personal circumstances, what would look, on the face of it, to be the most attractive?’. We argue a commons approach has very strong attractions. We can also say there are issues with whether or not that’s feasible, so we address that issue as well.
Daniel Susskind: One way to think about the professions we have is that they are gatekeepers — we ask each of the professions to look after, curate and update their own respective bodies of knowledge. Doctors look after medical knowledge, lawyers look after legal knowledge, tax accountants look after knowledge of how to fill out tax returns and so on. One way to think about what’s happening at the moment is that we’re seeing the decline of these old gatekeepers.
So the promise of all of this is the liberation of the practical expertise that had previously just been locked up in the heads of professionals or been buried away in filing cabinets. Now, there’s the promise of the liberation of it. There’s also the threat of new gatekeepers, new organisations — perhaps internet companies, perhaps technology companies, perhaps search engine companies, data companies — who take control of these new bodies of practical expertise.
A great example is in the world of tax. In 2014, 48 million Americans used online tax preparation software rather than a traditional tax accountant. That’s a big deal in the US where most of the time their tax system changes — on average I think it’s once a day — so most people employ professional tax accountants to help them. I think about 35 million of them used software from TurboTax which is the biggest, most well-known brand.
Another question is, will all this technology take away everyone’s jobs? It’s question of timing — it just has to be the case that technology is displacing people from performing certain types of roles, but it’s giving rise to entirely new types of roles. It’s redeployment rather than unemployment, just people doing different things. The problem is that many of the new roles look very unfamiliar to traditional professionals, and many of them aren’t being performed by traditional professionals at all.
“It’s redeployment rather than unemployment, just people doing different things. The problem is that many of the new roles look very unfamiliar to traditional professionals, and many of them aren’t being performed by traditional professionals at all.”
One of the unhelpful things we do when we talk about the future of work is that we talk about the different jobs that people do. So in the professions we talk about doctors, nurses, teachers, accountants, but the term ‘job’ isn’t particularly illuminating because it encourages us to think of the work people are doing as kind of monolithic, invisible lumps of stuff. In actual fact, people do lots and lots of different activities in their jobs, lots and lots of different tasks.
So the most useful way to think about the future of work is not by thinking about jobs, but by thinking about tasks. Are there tasks that even in this long-term scenario simply can’t ever be performed by machines?
I think most people, when they think about what machines are capable of, have what we call a ‘Rubik’s cube conception’ of machine capability. So there is a British man, who, with LEGO and a smartphone, built a machine that could solve a Rubik’s cube in three and a half seconds. He did it at his dining room table over the course of six months. The world Rubik’s cube champion can do it in four and a half seconds, so he’s outperformed by LEGO and a smartphone.
Of course machines are great at routine tasks, but we do much more than just perform routine tasks, we perform non-routine tasks — we do things that require creativity, judgement and empathy, and [tasks] that require creativity, judgement and empathy will never be performed by machines. That makes people quite optimistic about the future, but I think there’s two mistakes being made there.
The first relates to the decomposition idea — when you break down the work that most people do, it turns out that not that much actually requires creativity, judgement and empathy. People tend to overstate the non-routine component of the work that they do, and underestimate the routine component.
The second thing is that it’s a mistake to also think that the non-routine stuff also can’t be performed by machines.
Richard Susskind: Many people say, ‘As an expert I use my judgement, but how can a computer system ever exercise judgement?’. We say that’s asking the wrong question, the question you should ask is: ‘To what problem is human judgement the solution? Why is it that we need human beings to exercise judgement?’ — that’s the fundamental problem.
Our take on that is we exercise human judgement under conditions of uncertainty and we live in a world of uncertainty — the facts are often uncertain, the knowledge that’s applicable is often uncertain — so we go to experts because they’re the best people at handling that kind of uncertainty.
So the bigger question is not, ‘Can we programme a computer to exercise judgement?’, it’s, ‘Can we develop systems that can actually handle uncertainty to the level of a higher standard than human experts?’. The answer to that is almost certainly yes, and that’s where the fields of big data and incredible processing power come together.
Daniel Susskind: People say [a non-routine task] requires creativity, requires judgement, requires empathy, and the temptation for everyone is to say, ‘Because a machine can’t think like a human being, it can’t be creative; because it can’t feel like a human being, it can’t be empathetic; because it can’t exercise reason like a human being, it can’t show judgement’.
The mistake we’re all making is to fail to realise that machines can perform these non-routine tasks that might require creativity, judgement or empathy from a human being, but they can do it in a very different, almost un-human way.
It’s an exciting or horrifying thing to think about depending on your standpoint. The worry about spending too much time thinking about a jobless future 150 years away is that we don’t concentrate on what these systems are capable of doing today.
“The worry about spending too much time thinking about a jobless future 150 years away is that we don’t concentrate on what these systems are capable of doing today.”
It’s not going to create massive unemployment, but it could transform the way that we solve really important problems in society. That, for me, is the thing I really want to focus on — how is it changing the things people do in their jobs, how can we make sure that people get the right skills and training in order to do those things really well, and how can we solve these problems that we’re not solving very well at the moment?
Richard Susskind:: I don’t frame it [as a societal threat] at all. When you think globally of seven billion people, how many people have actually access to medical help or to help about their entitlements? That, for us, is the social ill we’re trying to address.
The challenge is to see, ‘What jobs should we, now, as human beings, take on in working with these increasingly capable machines in sorting out some hitherto intractable problems?’. It’s tremendously exciting.
The treat is that ways of working that we’ve taken for granted for many years, may no longer be viable. We see that as a far lesser threat that the notion that when we could do otherwise, we’re still contemplating the notion of allowing our expertise in society to be held by gatekeepers.
If I can speak as a lawyer for a second, if these technologies give rise to loss, damage or injury, who’s responsible? There’s another question [of] the extent to which you’re negligent in not using some of these technologies. You may say [of] a medical diagnostics tool, ‘Oh I’m not going to use that because there’s not a human being involved’, if it’s overwhelmingly shown to deliver better results, you’re more likely to be responsible from a legal point of view for not using it than using it.
I speak a lot to judges about this because it’s a kind of evolving new set of circumstances. My own very strong feeling is that we shouldn’t be holding people responsible too quickly for defects in system design where the systems by and large are overwhelmingly better than human beings.
We don’t want to deter innovation; we don’t want to discourage people from developing systems that could really make us far better off for fear of there being a slip in one line of code that gives rise to a loss.
“We shouldn’t expect more from our machines than we get from our human beings. People say, ‘That’s a computer system doing that diagnostic, there could be an error in the software’, that’s true, but the level error in human conduct are remarkable.”
We shouldn’t expect more from our machines than we get from our human beings. People say, ‘That’s a computer system doing that diagnostic, there could be an error in the software’, that’s true, but the level error in human conduct are remarkable.
Daniel Susskind: There’s a robotic pharmacist at the University of San Francisco in California. No human being’s involved, it’s made six million prescriptions so far and it’s made one error. A lot of people would see that error and say, ‘This is appalling, unacceptable, start dismantling the robot, start suing…’. But you’ve got to remember that human pharmacists make errors at best 1% of the time — that’s 60,000 mistaken prescriptions rather than one. There is a risk that if we hold these systems to a higher standard than we expect of human beings that we kind of chill. It would be a tragedy if the driverless car possibility was extinguished by legal uncertainty.
The mistake is to be talking about jobs. Forget about jobs, ask what sort of tasks have to be done to perform those jobs. In the 19th Century the new jobs that were created required tasks very different to the ones required in agriculture. The question now is whether or not the things we can’t even conceive of, and they may be entirely new types of goods, genuinely require new types of tasks?
As machines get more and more capable at more and more types of tasks — whether it’s cognitive tasks, emotional tasks, physical tasks and moral tasks — it seems more and more likely that the new products or services we create will be able to be produced by the sorts of tasks that machines rather than people have the advantage in.
When people look at Google and Facebook and say, ‘Look at the value of these companies, but look at how few people they employ’, the observation that they’re making is that this new wave of growth is being driven by things that don’t necessarily require as many people to perform them.
We close the book by asking two big moral questions. One is, are there some tasks that even if machines could do, we simply don’t want them to do? The second is, who should own and control all this new stuff online?
“Are there some tasks that even if machines could do, we simply don’t want them to do? The second is, who should own and control all this new stuff online?”
Our view is very clearly that we want to put as much stuff as possible out there in a commons. That is a position that lots of people would disagree with, and I think most important thing to do is to recognise that the old gatekeepers are crumbling, new gatekeepers are rising up, this is something we need to debate and think about — it’s not for one or two of us to say this is the right direction. We need to have a conversation — more people need to be talking about exactly the sort of thing you’ve talked about.
It’ll be driven not by a few individuals, but a swell of feeling that this expertise, this ability to solve important problems, ought not to be controlled by a few individuals, but should be made more widely available.
It’s not just transport in London that’s owned by a private company, but it’s also knowledge about how to solve health and legal problems. These things are pretty significant parts of our lives.
Richard Susskind: For me, the future looks nothing like the past. I don’t think we should dwell too much in optimising and streamlining essentially antiquated professions. I think the real challenge is to use technology to allow us to reach people and deliver a higher quality of service, far more accessible, far more affordable, than was conceivable in the past. We’re not going to do that by piggybacking on 20th Century technologies, we need to rethink.
This talk took place st Second Home, a creative workspace and cultural venue, bringing together diverse industries, disciplines and social businesses. Click here to find out who’s speaking next.