Can AI Reduce the Burden of Human Capital Investment?

Prior to the deep learning which is increasingly associated with AI, the progression of knowledge required ever more personal time commitments (in aggregate) on our part. Granted, AI has its own costs, as it would require extensive electrical capacity in the future. Even so, electrical resources could supplant extensive time requirements for high skill credentialing, for a mere fraction of today's human capital investment costs. Of course, the working terms of potential substitution are presently far from clear, even as pleadings for human-centric knowledge application are already underway.

However, there's an important aspect of this development which has yet to be taken into account. Only recall that (in part) due to human capital investment requirements, societies began to categorize skills use and labour divisions into ever hardening class structures. For that matter, lower income levels are increasingly being discouraged, from taking on the "necessary" human capital investment requirements of today's premier workplaces. The fact many workplaces are no longer capable of compensating formal educational costs, even as taxpayers remain on the hook for doing so, doubtless contributes to this gloomy dialogue as well.

The good news is that due to the low operational costs of AI deep learning, decades of human capital investment for high skill knowledge application are becoming less necessary. In other words: We no longer have suitable rationale to insist that strict divisions of knowledge based "labour" are "necessary" due to personal differences in income and resource capacity. The not so good news, however, is that societies may elect to not take advantage of AI's ability to reduce present day burdens of human capital investment - thereby keeping unnecessary class divisions in place. Should society choose the "not so good" option, the production and consumption of many important time based services, could eventually devolve mostly to higher income levels. As a productivity issue, one might frame such a choice as preserving excess inputs in relation to time based services outputs, so to preserve the "dignity" of our elite.

Yet AI could restore hope and economic progress, via its immense capacity to contribute to deep learning. By working with AI instead of in spite of it, most of us would become able to make knowledge and skill count in the immediate contexts of our daily lives. One benefit of working with knowledge on an "as needed" basis, is that this organizational process encourages the moderate skill levels so many individuals possess. Further, moderate skill levels are particularly well suited for the interdisciplinary approaches now required for problem solving. Options such as these are particularly important, as the half-life of knowledge can sometimes impact what we learn via traditional or formal educational settings. As Shane Parrish recently noted:

As a body of knowledge doubles so does the cost of wrapping your head around what we already know. This cost is the burden of knowledge. To be the best today in a general field requires that you know more than the person who was the best only 20 years ago.

Imagine being able to apply much needed knowledge and skill for a wide range of economic activity, in a fraction of the time presently required. In "Whiplash: How to Survive Our Faster Future" by Joi Ito and Jeff Howe, the authors describe their own extensive experimental efforts for learning processes. In the chapter "Practice over Theory" they write:

Putting practice over theory means recognizing that in a faster future, in which change has become a new constant, there is often a higher cost to waiting and planning than there is to doing, and then improvising. In the good, old, slow days planning - of almost any endeavor, but certainly one that required capital investment - was an essential step in avoiding a failure that might bring on financial woe and social stigma. In the network era however, well-led companies have embraced, even encouraged failure. Now, launching anything from a new line of shoes to your own consulting practice has dropped dramatically in price, and businesses commonly regard "failure" as a bargain-priced learning opportunity.

In time arbitrage, where learning would be integrated with workplaces as a time continuum, AI could assist individuals and groups in knowledge coordination processes, instead of replacing human capital components just because it seems "expedient" to do so. Should we give ourselves permission to fully participate in knowledge use, AI could also be given permission to relieve the burdens of human capital investment. The result might just be a new frontier, in which we can scarcely imagine the possibilities.