This is the English version of my weekly column for Forbes.
The original French version is here
This time is different. This is often how the impact of innovation on the economy and jobs is perceived. Should we banish the broadly deterministic analysis I made in my last column (available here) or is this still the right approach? In other words, are innovation and employment complementary when all is said and done (this is the lesson that history teaches us according to Alfred Sauvy, and Paul Krugman’s model suggests the same interpretation), or is there a danger that the impact of innovation will be negative for jobs in the long term?
The “this time is different” theory does not work in a number of situations and so we need to remain cautious on this type of reasoning. Carmen Reinhart and Kenneth Rogoff wrote a remarkable book on the application of this notion to the financial crisis (“This time is different – Eight centuries of financial folly” Princeton University Press 2009). They suggest that during the various financial crises, “this time” is never really different even if spectators get caught up in the moment and view each crisis as a historical turning point at the time.
But the innovation that we are seeing in the economy today reflects a radical shift for the way economies operate. It marks long-term deep-rooted change and not just the usual frenzy we have traditionally seen during each financial crisis. There can be no going back to the situation we knew before (which is precisely what a financial crisis is), and this is exactly why the wealth of innovations could give a different end result to the scenario we may have expected on the basis of past experience.
Looking at another aspect, innovation affects all areas of personal lives and professional careers. Perhaps it is new technologies’ very ability to cover all aspects of our lives that generates the most uncertainty and raises the most questions.
The way various technological innovations are examined below is heuristic. There is no notion of comprehensiveness, and there are also strong interactions between these various factors. However, we note that in one way or another, these innovations push back technological limitations and therefore trigger a fundamental shift in the arbitrage between capital and work.
Artificial intelligence eliminates Polanyi’s paradox; machines can learn even when the precise steps involved in learning have not been explained by a human being. They can therefore make appropriate decisions, and concrete examples of this include the technological application in our smartphones (e.g. Siri on the iPhone) and in a number of testing processes (medical, legal, etc.).
Machine learning provides analysis and decision-making capabilities that are more detailed and precise than anything human beings are capable of. Artificial intelligence combines with Big Data to revolutionize decision-making processes and could exceed humankind’s analysis and decision-making abilities. This paradigm shift changes the entire landscape.
This shift is even more striking when we see how artificial intelligence can learn by itself and create its individual momentum that is truly its own. The job required to set this process in motion is then no longer required. What about computer programmers, who are so vital and key to this innovative technology…what will they do when machines are able to efficiently program themselves?
Robots can manage complex operations with virtually zero error, and they can now replace humans on a wide range of tasks, not merely in the industrial sector. The services sector is also affected, with a good example being the Fintech sector, which operates using considerable robot technology.
Robotization no longer merely involves arduous and repetitive industrial processes, but also stretches to the services sector. This may generate high productivity gains and cut back jobs, but where will these jobs go? What sectors will provide work and help re-establish the macroeconomic balance of full employment?
Algorithms can very swiftly resolve complex questions, as shown by high frequency trading. Computers make calculations at a speed that man is incapable of, so human beings are being replaced as machines are too fast for them. A recent article (March 7) in French financial daily Les Echos reported that an investment bank had cut back its team of market traders from 600 to two but recruited 200 IT specialists instead, in a cheaper and more efficient strategy.
Machine-to-machine communication: a machine sends a signal instructing another machine to carry out a specific task without human intervention. This can be very useful in a medical context, for example: a sensor attached to a patient can send a signal after monitoring one or several vital signs that track the patient’s progress, and the machine receiving the signal can adjust the patient’s treatment accordingly.
These decision-making processes can now take place without human intervention.
3D printing technology involves printing an item in three dimensions with great ease. It also allows for remote production management. Rather than producing goods in one place and having them transported, it can be easier and more straightforward to install the 3D printers in the destination location and manage the process remotely from the initial location. This profoundly alters the entire production process and makes for a very interesting situation as we can envisage a European or American company setting up this type of process in an emerging country, and conversely an emerging market-based company managing the process remotely in a developed country.
Autonomous vehicles, both cars and drones, are used without human intervention and with very promising results. Transport is a major source of employment, especially in the delivery services sector, and employs drivers who often do not hold many qualifications and can be seen as the equivalent of yesterday’s blue collar factory workers. Driverless cars mark a massive and definitive shift.
We should also mention the internet platforms that are transforming the way we live and behave. Blablacar, Amazon and Airbnb are revolutionizing consumer behavior by offering an original product or service. These companies are enjoying increasing returns, and their size is therefore improving their productivity and profitability. Business has so far operated with diminishing returns: higher production has traditionally meant lower profitability. Yet these platforms have an entirely contrasting business model: size has become an advantage as it reduces marginal cost, and this new model entirely changes the way we analyze competition. These companies’ momentum is derived from computer networks and the use of artificial intelligence and robots. These very powerful companies do not actually employ many staff. The size of a company is therefore no longer a guarantee that it will provide a large number of jobs.
All these changes reflect how machines are now influencing the way jobs are carried out, whereas until now, the machine has always simply replaced the job. This situation is set to become even more complex as artificial intelligence can already create its own dynamic without human intervention.  
This massive transformation is at the very heart of all questions on employment. There will obviously be a negative long-term impact on jobs, for four reasons:
The first factor is that even if we want to agree with Alfred Sauvy, adjustment of the labor market will take a long time.
All business sectors will be affected. We will not be able to get around sectors that are not affected by these innovations. When we refer to the Luddites in England or the Canuts in France, we are talking about a specific business sector and the issue of its production methods, but digital transformation is now the buzzword on everyone’s lips across all business sectors, so moving from one sector to another is not necessarily the answer that it was in the past.
This is only the start of all these trends, and companies are only just beginning to fully roll out a real digital approach.
For example, banks in France are just now beginning to adopt real digital strategies. A recent report by McKinsey confirms this situation and suggests that spectacular transformation lies ahead: perhaps just like the last ten years or at an even faster pace (the smartphone is only 10 years old, how did we manage before?). Transformation is going to continue, and so the world of tomorrow does not exist yet. This means that if we want to follow Sauvy’s logic, it will take years or even decades to strike a new balance. Historians 500 years from now might tell us as much, but it will be too late for us by then.
The fourth point is the issue discussed by Richard Baldwin in his latest book (“The Great Convergence: Information Technology and the New Globalization” Harvard UP November 2016). Technological progress will lead to remote operations on a large scale, similar to our example on 3D printing but with a much vaster scope. Globalization will no longer be about cutting transportation costs, but will mean the ability to manage operations remotely via a complex digital service. This will lead to even fiercer competition of many dimensions, which means that companies will be under greater pressure to remain at the cutting edge of innovation, with the ensuing risk that jobs will be hard hit.
It is impossible to assess the end effects all this will have on jobs. Complementarity between jobs and innovation can win out and take root in the long term, but in the short term, until such times as the situation stabilizes, this machine momentum is likely to win out over human jobs.
Economists have however endeavored to assess the impact of these innovations on the current labor market. A first study was published in 2013 by two economists at Oxford University. If we look at jobs facing direct competition from machines, the authors estimate that 47% of posts in the US are potentially under threat of being replaced by machines in the next 20 years. This methodology met with criticism, as an individual worker in one single job carries out a number of different tasks that are not all rivalled by a machine to the same extent. The OECD broke down jobs into individual tasks and estimated that around 10% of jobs are under threat from machines. The OECD’s figure is reassuring, while the Oxford University research is worrying.
These figures are deterministic, yet the economy of the next 10 to 20 years has still to be invented. In view of the technological shifts we are seeing, I think that 100% of jobs are at risk. All jobs will be affected either directly or indirectly by these innovations.
These various developments reveal the need for substantial efforts on education, restricting the risk of excessive divergence between the development of machines and man’s ability to take ownership of this technological innovation. Technological progress is marching on at high speed and the labor market must adapt to meet this new context: herein lies the real challenge.

We are building a somewhat crazed society: work is the way we define who we are, yet it is slipping out of our grasp and becoming dictated by machines. This situation is not going to come to an end by itself. The United States, Japan and Europe have worked hard on these issues, but China is now at the heart of these technological developments. Everyone wants to lead the race and this technological innovation is going to be at the center of competition.
We do not know how long this technological focus will continue to drive the spectacular progress we are currently witnessing. So we must prepare for an economy where very different ways of operating will have to exist side-by-side, one group that has kept pace with machines and another that has run out of steam. This is the complexity we need to get to grips with. And this is why this time probably will be different.