Tacit knowledge, a term coined by Michael Polanyi on his 1996 work The Tacit Dimension, refers to the kind of knowledge that manifests whenever we tacitly understand how to do something but are not then able to explicitly articulate it in terms of a procedure. In simple words, Polanyi’s Paradox states that ‘we know more than we can tell’. Polanyi’s argued that creative acts (especially acts of discovery), although possibly aimed at discovering the truth, are charged with strong personal feelings and cannot therefore be stated in formal terms. Examples include making a persuasive argument, recognizing people from a distance or recognizing someone as they grow up.
Tacit knowledge poses an extreme challenge for computerization and automation, as they both work by first taking the explicit procedure and then codifying those steps so that a machine can do it in our place. The fact that high-level reasoning requires very little computation, whereas low-level sensorimotor skills demand enormous computational resources is in fact a well-established paradox in artificial intelligence research.
Now, what does this tell us about the future of work? Will we have machines that will do what we do, programmed to mimic this kind of nonprocedural human way of assessing and solving a problem? Or will we recast the problems and therefore the technologies, so that machines do it very differently from how humans do it and yet still successfully? And either way, will technology ultimately replace human employment?
Think of the work of an administrative staffer or an assembly line worker – the type of skilled work that uses a lot of procedural – and therefore easily substituted by machinery – activities. These positions are likely to be almost fully substituted by machines. Conversely, a scientist or an attorney make use of their creativity, intuition and expertise to perform their tasks. These jobs, in which a lot of technical (codifiable) knowledge is also required, are complemented by information technology (i.e., it allows them to be faster and hence more efficient), rather than directly substituted by it. Finally, there are the in-person service jobs, like cleaning or personal care providers. As these require tasks that have proven very challenging to automate, they are expected to experience a relative growth in the future. But the irony here is that the supply of workers who can do them is quite abundant, which will push the wage level of relatively low-wage low-education jobs further down. This gives rise to the polarization phenomenon, where we have a simultaneous growth of both high-education, high-wage jobs, and relatively low-education, low-wage jobs. At the upper ends of the skill distribution, technology enhances one’s productivity; at the lower end, not so much.
To end on a positive note, I personally don’t think that polarization will go on until the middle just collapses to zero. For one, we can’t yet fully comprehend or imagine the ultimate types of human-machine complementarity. In the future, we might be using our time to do types of works we can’t imagine being possible right now or simply more artistic, less codifiable ones. However, although labor polarization due to increased use of technology does not create a problem of lack of wealth, it does pose one of income distribution: if the diminishment in labor scarcity among some skill groups should mean that their claim on the wealth is very limited, the whole capitalism system, which is organized around labor scarcity, will have to be rethought and reformed, so that society can be reorganized in a way in which people still have purpose that give meaning to what they do.