Somebody estimated that 60% of the five-year-olds of today would work in professions that do not exist yet. Learning a profession today, we no longer have a guarantee that it will survive during our whole work-life. The digital world that develops exponentially makes the job market change at a quick pace. It is clear today that the work division of today and the future is between humans, algorithms, and robots. The constant changes in AT learning and robotics are that quick that we should not expect any fixed equilibrium for this division at any time soon. The equilibrium will simply continuously shift, forcing people to continually retrain and upskill throughout their lives.
ALGORITHMS vs. WHITE COLLARS
The white-collar work got less annoying in recent years. Emerging of the excel sheet was a revolution that cut down many working hours in different professions. Later, with IT systems supporting the management of single enterprises and entire value chains, decision making changed the dimension. But today, we need to distinguish between programmable algorithms where machines make calculations designed by humans and artificial intelligence where deep learning schemes had stimulated the machines to find solutions earlier not possible to grasp with a human brain.
The bookkeeping and accounting rules are just mathematical functions. We deal with variables, parameters, and their interdependence. A computer does pretty much the same work that a skilled analyst would do. In the case of repeatable calculations, a well-designed excel sheet with some visual basic functionalities would already do. With tailored software that automatically collects and injects inputs, the work gets easier and easier. Only results are subject to qualitative interpretations. Do we rely on the contractual rolling stock as it is cheaper, or for safety reasons, do we want to keep a part of it as proprietary or leased? It is a kind of decision that has to be made upon the type of experience that a machine cannot have or acquire. With all functions programmed into the computers, algorithms will do the work for us, being more accurate, with no errors, and in a shorter time. What we earn from that? We have more time to analyze and interpret. But is it any kind of a true artificial intelligence that makes all the work instead of us? No, I do not think so. It is just computation. The value-added delivered by the technology is limited in this case. It is rather time savings, including those on looking for errors. The invention of the excel sheet some time ago made the accounting services cheaper and hence, affordable for anybody. Is there space for a new dimension of accounting, or will the working hours shift indeed from accountants to algorithms in the future with accountants barely having the idea of what other value-added could they contribute to the managerial process?
In the business world, the needs for problem-solving are often of a more complex nature than in accounting. In logistics, for example, freight management systems require a full recalculation of flows made on a daily basis or even real-time recalculations. It is not only the accuracy of calculations that counts. It is the time constraint that matters. The quicker the injection of variable inputs and the faster the recalculation with all the interdependent functions, the bigger the serviced network can be. The algorithms and open-source computing power allowed the logistics companies to build more complex networks with more consignments served daily at the lowest cost possible. The value-added delivered by digital technology is much higher than in the accounting business. It allows for economies of scale on unprecedented volumes. The delivery network designed by algorithms becomes a digital twin of the real delivery network. The digital twin, who is the firstborn one, is the virtual twin. The other twin is its reflection in the material world. For the last thirty years, the digital twins disrupted the traditional transport industry, making way for the logistics industry that, in turn, got to one of the major employers. White collars in logistics today are those who interconnect with their clients often on a personal level, and the job requires even more creativity. Could algorithms take over here, as well?
Algorithms up to recently took over only the simple tasks that were programmable and for a longer time, might have turned to be quite borrowing for a human. Unless, of course, somebody likes this kind of work. I saw one or two exceptions in my professional life. But I am convinced that the rule that everybody who mastered the basics ultimately needs a challenge and value-added autonomous and more creative work still applies. And with algorithms performing the basic, repeatable work for us, jobs are getting indeed less borrowing by requiring more. More can be challenging. Professional burning out of the boredom will, for sure, not apply in the future. But being always creative, we might burn out by overheating … A quiet hobby would provide a balance.
Or, will we need somebody with whom we could just talk and release stress; discuss our life and possibilities? The demand for psychologists and coaches is on the rise. Up to recently, I had no idea how much. Till out of the blue for some diversion, I went through postgraduate studies in coaching and mentoring. And as a boss, I have learned to listen and observe. Besides decision making, the other competence that matters much in the present-day world.
With advances in artificial intelligence teaching or learning, the white-collar support gets even stronger. What if an algorithm can recognize the early stage of illness on a picture generated by the ultrasound scanner, upon shapes that would be not relevant for a human eye? It is not a function that had been injected into the computer. These were thousands and thousands of pictures with a diagnosis made by human doctors either upon a picture at sight or later as the illness advanced and was recognizable, or even after surgery had revealed the problem. All the material was fed into the machines. The machines have learned on their own how to recognize the illness at an early stage. The doctors’ work was augmented the way that it is doctors’ who can learn the diagnostics from the machines. The artificial intelligence delivers even higher value-added. In fact, it augments the doctors, work. Will it replace him or her? I do not think so. But a consultant whose advice is highly reliable would make the doctor’s job less stressful for sure.
ROBOTS vs. BLUE COLLARS
During the industrial revolutions of XIX and XX century, the steam and later the electrical power allowed to mechanize some physical work. With the implementation of robotics, the XXI century is marked by automation. Unlike in the case of algorithms, the advantages of mechanization and automation is rather not the kind of the value-added that can be delivered to the white collars. It is continuously taking off the hardships of physical or dangerous tasks.
The robotics era began en mass with the implementation of industrial robots that were caged for safety reasons. They work quick. But they are only fully programmed. Making their programmed moves, they would not see a human approaching. Anybody approaching them at the close is endangering him or herself. Although dangerous at the close, industrial robots took over much physical work from people. This was not necessarily tasks that were the hardest ones. The early stage of movement control and sensor or visual control allowed only employing robots to deliver programmable and repeatable physical work. The kind of work that normally incurred work-related diseases. It is mainly the production lines in the automotive and the electronics industry. Robots just took over. The blue-collar jobs were replaced but not augmented like in the case of doctors getting support in the diagnostics field. Redundancies had been the result. From time to time we hear that a leading car manufacturer wants to cut jobs at the production lines. Recently redundancies going into thousands were announced by one of the German automotive corporations.
But, new white-collar jobs have been created to program and maintain the robots instead. The job market increased. But like accounting or financial experts, the automotive or electronics industry workers have to look for another job. But it is not only changing the job. Often, it is changing the job profile.
The other major utility of robots is to assist humans in some difficult and dangerous tasks. In the latter case, the service robots are not fully programmed and automated but instead remotely controlled by people using them. The Notre Dame cathedral fire rescue had shown to the whole world how a remotely controlled robot can successfully take over. Without Colossus (it is the robot’s name), who poured lots of water standing midst of the burning oak and stone falling down to lower the inside temperatures, the structural damage to the cathedral would be much more significant if not destructive >>>.
The sensor, visual, movement, and gripping control technologies advance as well as the AI training techniques. The industrial and service robotics gradually gets to yet another dimension, where jobs are no longer replaced. They get augmented.
The currently emerging sector of robotics is collaborative robotics. Collaborative robots or ‘cobots’ and humans will be co-workers performing work together, sharing the same working space. The robot will not directly jump into somebody’s position. It will assist her or him in doing their job performing tasks requiring among others but by far not only physical strength. White-collars work with laptops filled with algorithms. Blue-collars will be assisted by corobots. Something like Luke Skywalker and R2D2. 3PO was rather a protocol robot or in present-day language personal service robot. For the latter, we will have to wait a bit longer …
Cobots will share the working space with human workers taking over only the repeatable and heavy tasks. The ultimate target is to teach machines (which is possible only with advanced machine learning) synchronize with their human co-workers, preferably in real-time. Today the concept works already in logistics where cobots (that, in fact, are automated trolleys) follow a picker in a warehouse. Movement alongside the warehouse shelves is predictable; hence, doable for the present-day programming. It gets more complicated with gripping, as the human movements are not predictable. Too many human moves by a human co-worker would make the cobot repeat an error protocol over and over again. The alternative way of automating picking is, by the way, to automatically bring goods to the picker. But ultimately, it is the picker who makes the consignments. Some human work is simply not replaceable. At least not yet and seemingly not for the nearest future.
The unpredictability seems to be the major obstacle in developing automated robotics. The most prominent example is the development of automated cars. Its advancement stalls at unpredictable traffic conditions between free-flow and synchronized flow >>>. The other major obstacle is the level of sensorimotor skills like flexible movements and grips. Polishing and grinding that require continuous fine-tuning of pressure applied to the surface are still beyond the competencies of robots of today.
Overtaking those thresholds will require significant work in the field of AI as well as vast computing and transmission power. The latter is needed so that machines via a cloud exchange their experience. AI learning depends on the collection of samples (big data) that are used by the AI to master the skills. The bigger the sample of possible cases, the higher the predictability of the process for a machine. As machines can learn today by repeating the human moves, the sky seems to be limit here. Collecting samples, bundling similar cases together, and turning unpredictability into patterns requires, however, time as well as transmitting & computation capabilities. We will have to wait a bit yet.
But still, the less repeatable the working day on a job is, the lower the probability is that a machine would take over in the nearest future. (to be continued)
Photos by: Vitaly Vlaslov, Quintin Gelar, LinkedIn Sales Navigator, Kateryna Babaieva, Kevin Ku, Alex Knight, Pixabay