Artist Jorge Otero Escobar weaving Lomo (Backbone) in 2015

The Future of Work

Artist Jorge Otero Escobar weaving Lomo (Backbone) in 2015

I just read a blog post, The Future of Work – Redux by John Philpin. It provides a nice, short look at what might happen as computers, robots and artificial intelligence become increasingly present in the workplace—what will people do when “all the work is done by robots?” As a result, I will be using computer, robots and AI interchangeably for the rest of this post.

John expresses a view that the future includes people working with robots, not simply people being replaced by robots. I happen to agree with that. I’ve written several blog posts on artificial intelligence (AI) and my skepticism about the capabilities and pace of the introduction of AI systems. AI has enormous potential, but I don’t see AI making humans obsolete any time soon (actually, I don’t see AI making humans obsolete—period).

Computers, and by extension, robots and AI, possess one important capability: they can add and subtract really friggin’ fast. George Boole developed what we now call Boolean Logic and it created an approach that allows us, following in the footsteps of Charles Babbage, Augusta Ada King-Noel Countess Lovelace (nee Byron), Grace Hopper and Claude Shannon, to stick those additions and subtractions together in such a way as to resolve any computable task (à la Alan Turing).

Since we don’t really understand what intelligence is and what generates intelligence, it’s difficult for us to create a tool that mimics human capabilities. That does not mean that we cannot use AI in some kind of evolutionary process to develop increasingly higher performing AI—only that it isn’t likely to mimic (and by extension replace) human intelligence. We will need to understand what those differences are and how computers / robots / AI and humans can work together synergistically.

These insights in hand, I want to look at what might happen in the near term as more and more capable AI systems make an appearance in the workplace. I think that there are a limited number of ways it will impact the workforce:

  • AI systems will replace one or more humans completely. I don’t see this happening any time in the near future for jobs requiring significant cognitive skills. Such human jobs tend to encompass a great deal of implicit diversity of tasks. For example, an administrative assistant’s job of making travel plans and reservations could be done by an intelligent agent. It’s not likely that said intelligent agent can then build the travel packet to accompany the traveler. I know that the traveler can have all that information downloaded onto a portable device to accompany them. However, to quote a deployed US Marine, “A computer with a bullet hole in it is a brick. A paper map with a hole in it is still a map.” While it’s unlikely that anyone will shoot a hole in the traveler’s digital device, there are any number of minor problems that would become a catastrophe should that device be impaired (which I know never happens—except, per Mr. Murphy and his eponymous law, any time that it can).
  • AI, or perhaps more appropriately automation, will replace jobs in which there is a limited requirement for cognitive skills. Jobs have already been replaced by ATMs, and by touchpads in some restaurants. This is very much akin to the replacement of agricultural workers by mechanical reapers, threshers, and balers in the nineteenth century, and is a pure productivity enhancement.
  • AI systems take over part of the human’s job. This is akin to the industrial case where power tools replaced hand tools in factories in the early nineteenth century. It makes the human more productive and organizational leadership then has the option of how to use that productivity.
  • AI systems start doing a job that wasn’t done before. This is the Big Data job. By definition, analysis of massive data files is not something that was being done by humans. This can result in an increase in productivity of the organization, but it may manifest an advantage in other areas than productivity—areas such as exploiting market niches previously unrecognized. It’s possible that the results of this analysis could lead to the need for additional human jobs.

If I am correct that the introduction of AI will result in an increase in productivity, organizational leadership can view themselves as suddenly having additional resources (the time of the people freed up as AI systems do some of their work) to apply to previously un-resourced activities. Every organization has such un-resourced activities—things that the organization would like to do, but for which it doesn’t have the resources. A company may have wished to develop two new products but had man-hours for only one and now may be able to accelerate development of the second product. Alternatively, a non-profit may now be able to provide services to additional clients.

The application of AI systems can help redirect those resources.

The challenge is for leadership to make the proper fit of human resources to those previously un-resourced activities. Just because the use of AI freed up four hours of skilled time a day doesn’t mean that leadership could assign a machinist to create computer graphics.

Admittedly, this is a “glass half full” perspective, but we, as humans, have a terrible track record of predicting the labor impacts of the introduction of new technologies. The industrial revolution didn’t result in massive factory unemployment, it merely changed the kinds of jobs the workers did. The introduction of personal computers into the workplace thirty years ago didn’t result in massive unemployment among office workers, it merely resulted in the gargantuan increase in the number of PowerPoint presentations. I suspect that AI systems will be a factor in changing a number of professions that require significant human intellectual effort but are now being done in a mass-production fashion. My exemplar would be teaching. Today, teachers almost never have the time needed to provide the personal, one-on-one attention needed to identify and develop individual talents, and people are individuals with different abilities and different learning styles. Mayhaps humans liberated from the drudgery of office work can become part of an educational workforce that leads to a future in which we approach a situation where every person actually has the opportunity to gain a quality, personal education contributing to the betterment of humanity.