When giving a talk to macro economists, one the surest laugh-lines is to claim: “This time is different”. The merriment arises from the fact that many of the biggest follies in forecasting were founded on the belief that the usual rules have been repealed – that the normal constraint on asset values, growth rates, or whatever projection don’t apply this time.
It is thus with eyes wide open that I claim that “this time is different” when it comes to the future of work. Not everything is different, but many important trends are – in my view – significantly underappreciated. Let me start by explaining my basic premise.
Globalisation + robotics = Globotics
(i) This phase of globalisation and robotics (or ‘globotics’) will mostly be about the service sector; not just the manufacturing, mining and agricultural sectors as in past decades. This matters in many ways – not the least of which is that 80% to 90% of people in advanced economies now work in the services sector. The implications are clear. If you are judging the social impact of future globalisation that affects the majority of jobs by comparing it with the impact that past globalisation had on a much smaller share of people who work in the production and extraction industries, you are underestimating the number of people affected by the gains and pains of globotics.
(ii) Globotics is governed by a different physics. Past globalisation and automation were mostly about goods, and thus the manufacturing sector and related infrastructures. They were therefore ultimately restrained by the laws of physics that apply to tangible things. The globalisation and automation of the services sector are all about information (electrons and photons); processing it and transmitting it. Globotics is thus ultimately linked to the laws of physics that apply not to tangible things, but to electrons and photons. This alters the possibilities.
It would be physically impossible to double world trade flows in 18 months. The infrastructure could not handle it, and building infrastructure takes years, not months, for reasons that have to do with physics as well as economics. World information flows, by contrast, have doubled every couple of years for decades and will continue to do so for years to come. This timescale disparity is due to the physics involved.
Electrons can ignore many of the laws of physics that slow down globalisation and automation in industry and agriculture. The technological impulse behind the incoming globotics transformation is profoundly different to the technological impulses that triggered previous waves of automation and globalisation.
The implication of this is clear: if you are judging the pace of globotics based on the pace of past globalisation and automation, you are using the wrong physics. The background constraints that shape the economic pace of change have changed. This is why historical experience must be treated with great care when applying lessons to today’s globalisation and robotisation. And it is exactly why the disordering of services sector jobs will come faster than most believe.
We live in a time when the future of work is changing rapidly, but also in a time when it is particularly difficult to sort out the main lines. The baseline argument is that the world we are in is witnessing two distinct types of growth processes. Let me label them “expected” and “explosive”.
- The first is the normal or “expected” pace of the post WWII growth process, where 5% to 10% per year is a really fast growth rate for just about anything ranging from GDP and sales to wages and productivity. The doubling times for these growth processes are measured in decades.
- The second is the “explosive” pace of almost anything to do with digital technology – which means the collection, transmission, storage and processing of information. The doubling time for these growth processes can be measured in months.
The tricky part is that most things – say job displacement, or reorganisation of work practices – are affected by a blend of the two processes. The difficulty lies in sorting out which bits are subject to expected change and which to explosive change. Take the example of a patent lawyer in California.
James Yoon has a great job as a lawyer specialising in patent disputes. His job today is very different to what it was 20 years ago, but in other ways it is exactly the same. Today, he charges clients USD 1 100 an hour – much more than the USD 400/hr he charged in 1999. This is not just because he is more experienced. It is also because AI-trained computer programs have transformed some aspects of his job. Today, he organises a team with a quarter of the legal human manpower since he is using white-collar robots. These robo-lawyers are good at things like searching through documents and emails, and flagging which ones will be relevant. They have helped Yoon to cut his cost base.
Specifically, he uses Lex Machina and Ravel Law to help him digest huge piles of court decisions and the documents filed on similar cases by the judges and opposing lawyers.
Nonetheless, much of Yoon’s job is untouched by digitech. He is still doing the most human aspects of the job – advising and reassuring clients, structuring the final legal strategy, negotiating with opposing counsel, appearing in court, and the like.
Artificial intelligence (AI) software robots – a tool not a replacement
More generally, AI software robots are making service sector workers and professionals more productive by automating some tasks – but not all of them. As a result, some aspects of the future are governed by the expected pace of progress and others by the explosive pace of digitech. The big pay-off in insight comes from thinking hard about which tasks are subject to the explosive progress and how this will affect the new jobs and those that survive the explosive progress.
One particularly common line of thinking takes AI-trained ‘robots’ as humans who haven’t quite grown into fully capable human replacements, but will eventually. The question posed is usually: how long will it take before AI completely replaces lawyers? I think this is the wrong way to frame the issue: it is more useful to think of AI as being like a tractor. A tractor is not a baby farmer that will grow up into being able to fully replace the farmer. A tractor is a very useful tool; it changed the occupation of farming and allowed fewer farmers to grow the same amount of food, but it will never replace farmers. AI should be thought of in this way for the foreseeable future. Robo-lawyers may eventually replace lawyers, but that is in the realm of science fiction, and not practical thinking ahead for today’s realities.
Telemigration – the future of globalisation
I have not touched much so far on the future of globalisation, but that too is changing rapidly and in ways few expect. The term I like to use is telemigration – people sitting in one nation and working in offices in another.
The first and most obvious linkage between digital technology and globalisation is automatic language translation – what experts call ‘machine translation’. Given that language barriers have been so important in separating people and hindering commerce among them, machine translation will change an incredible number of things.
Machine translation used to be a joke. Just two years ago it was little more than a party trick, or a very rough first draft – but no longer. Now it is rivalling average human translation for popular language pairs.
According to Google research, which uses humans to score machine translations on a scale from zero (complete nonsense) to six (perfect), in 2015, Google Translate got a grade of 3.6 – far worse than the average human translator which gets scores like 5.1. After a massive upgrade in 2016, Google Translate now hits numbers like 5.. The capabilities are advancing in leaps and bounds.
And machine translation is not something that is lurking in computer laboratories. It is on your smart phone, laptop, and tablet today. Free apps like Google Translate and iTranslate Voice are now quite good across the major language pairs. Other smartphone apps including SayHi and WayGo and are widely used. Google, for example, does one billion translations a day for online users.
Try it out. It works on any smartphone, tablet or laptop. Just open up a foreign language web site and apply Google Translate to the text. You can even use the iTranslate app to instantly translate foreign language in real time. You fire up the app on your smart phone and point your phone’s camera it at a page of, say, French and you see the English translation on your phone’s screen instantly and at no cost.
Another key change is the rise of freelancing platforms like Upwork.com. I like to think of these as the ‘container ships’ of future globalisation. They are how companies and individuals in rich nations are finding, hiring, managing and paying talented, low-cost service workers sitting abroad.
Conclusion: the future will be very different…
There is no proper way to end an essay like this one. Rather, it is meant as a call for more and more nuanced thinking about the future of work. But since it has to end somehow, I’ll finish with two of my favourite sayings. The first is the adage I used to end my 2016 book, The Great Convergence: Information technology and the New Globalisation: “Things have changed so much that not even the future is what it used to be.” The second is a quote from a social media marketer, Graeme Wood: “Change has never happened this fast before, and it will never be this slow again.” There is still time to prepare, but now is the time to wake up and realise that the future really will be very different.
 Google’s Neural Machine Translation System: Bridging the Gap between Human and Machine Translation, Yonghui Wu et al, Technical Report, 2016.
The Globotics Upheaval – Globalisation, Robotics and the Future of Work by Richard Baldwin, published in the UK by Weidenfeld & Nicolson in 2019.