I have this discussion with my good friends all the time now. I predict the destruction of most work and they think I’m crazy, or at least an extremist in my views. But then I read all of the reports about AI developments and other academic thinkers talking about jobs being at risk and think that I am mainstream, so which is it? Am I a doomsday thinker, or is my vision of the future reasonable.
I’ve come to realize in the last few days that the answer to that question is rooted in time and language. When I discuss these issues I see them as SO transformative to society that I ignore the short run and even medium run effects.
In Microeconomics, the short run is defined as a period of time not long enough to allow change to certain economic conditions.
Medium run is NOT a technical term from economics, but I am using it to characterize the transition period from short run to long run.
Most arguments being made today are short run and maybe medium run arguments for why AI and automation will create jobs. I actually don’t disagree with many of these arguments. But they are only short run arguments and that short run timeframe is shrinking exponentially. Our ability to adapt AI and automation to the needs of the marketplace has become spectacular.
So with that backdrop, how does it help to understand the challenges automation presents the nature of work. I have previously defined “work” as the sum of intelligence and physical labor. All jobs can can be defined as a mix of those two variables. The ditch digger… mostly physical labor. The college professor… mostly intelligence. When you think about work in this way, you can see where automation and AI can be SO disruptive. For centuries, we have built tools and technology that aided either physical labor or intelligence. But with AI and robotics, we are fully REPLACING the human element in the equation of work. One can argue that the AI can’t handle the intelligence required for certain jobs or the robotics aren’t nimble and dexterous enough to accomplish a task, and I am fine with that in the SHORT RUN. We may not yet have developed the AI to replace THAT job. Or built the robot that can handle THAT task. But do you believe that we CANNOT build those? or maybe won’t? If so why? Because I don’t see any reasons that we will not, for nearly all jobs. In the SHORT RUN, it is difficult to know what tasks and challenges will be created when a new technology is introduced, so it will create jobs. These jobs will be fewer and fewer over time because it takes fewer and fewer people to solve these tasks and challenges than it did even 20 years ago.
The disruption to work of AI and automation cannot be compared to something like the development of the Internet and it is that example of technological change that I believe is confusing people. The Internet took information that was disparate and segmented all over the world and put it in the hands of everyone instantaneously. That technology create millions and millions of jobs because it broadened the amount of work that could be done. However we did not have the REPLACEMENT aspect that AI and Automation bring to the table today. Said differently, and its a little absurd because AI and automation don’t exist without the internet, but if the internet arrived as a new technology today with AI and automation already at full speed, then far fewer new jobs would have been created because the machines can handle those tasks better than humans can.
But if you believe we will achieve General AI, then even the human reaction to deal with new challenges gets eliminated. General AI can deal with the new challenges faster than any human. So that brings us to the LONG RUN. In the LONG RUN, the machine solution is ALWAYS superior to human involvement in “work”. It is smarter, faster, more productive, less error prone and operates 24/7. In the LONG RUN, new technology does NOT create new jobs, because the human is eliminated from the work equation as an inefficient input. We no longer provide intelligence or physical labor.
It can argued that AI and automation will never fully replace humans in the equation of work and I grant that there may be some areas that they cannot, even in the LONG RUN. Humans may prefer to deal with other humans (will you always be able to know the difference?). Certain work may include amorphous concepts like HOPE and FAITH that AI may not be able to replicate. But to think that the majority of work cannot be replicated by AI and automation is to deny the human innovative spirit and centuries of technological advances, simply because it is hard to imagine or a fearful circumstance. I believe that an honest assessment of the skills (intelligence and physical labor) needed to execute any job will result in an example of AI or automation already being developed for that task.
As the founder of a group called ForHumanity, this may seem like an odd perspective, but it is just this perspective that has driven me to action. This is going to happen in the LONG RUN and we are woefully prepared for it. ForHumanity will spend its efforts to make this transition as smooth and as positive as possible for our constituency — Humanity. We welcome your assistance… it’s going take a village.