Learning Machines
It's not when machines start taking your jobs that you have a problem; It's when they start taking your classes.
The simple, conventional wisdom that stretches back to the dawn of the industrial revolution courtesy of old Ned Ludd (the origin of the term "Luddite") is that automation would progressively put all of us out of a job, leaving the working class, whom at that time were being steadily kicked out of rural agriculture and sucked into urban factories, with no means of sustenance. The Luddites resisted this trend, insisting that machines be restricted for the sake of preserving the dignity and livelihood of workers.
Needless to say, they did not succeed. Today, our understanding of the economic effects of automation are better informed by a much more historical data. What we discovered was that, by improving workers' access to education, they can be re-trained with new, more sophisticated skillsets that allow them to take jobs in emerging labor markets which would otherwise lay dormant for want of skilled workers. This is the model our labor markets operate within to this day.
Despite the evidence supporting this model, those of us in the US are watching our fabled diamond-shaped class structure sag, mostly due the steady march of a generation increasingly unable to afford the luxury of post-secondary education. But this is only the beginning of our problem.
Despite the fact that we have a better, subtler understanding of the problem of automation does not mean that the Luddites were entirely wrong. They simply did not take into account something that humans were better at than the machines of the time: learning.
But as we look out, now more than ever before, on our burgeoning technological horizons, we see innovators working on ever more ambitious projects. Among these lofty goals is been the creation of a learning machine - one that can benefit from our knowledge and educational systems as much as, if not more than, the privileged human beings who are currently engaged with them.
To be sure, such a system would be a powerful tool for much good - for whomever has the privilege of putting it use.
Add to this extraordinary development the strong possibility that such systems are likely to be easy to replicate after they reach a desired degree of skill (if you'll excuse the pun), and labor markets of every kind, finally, and again, look like bleak places for ordinary people to attempt to survive and prosper.
When serfs needed something to turn to after being cut from their lands, the factories were there to provide opportunity (such as it was). When workers needed something to turn to after being cut from their jobs, the schools were there to provide opportunity.
What will be there for us when the schools, also, fall to automation?
I believe the answer will be democracy. But if that is to be true, then we will have to make it true. We have work to do - and not the kind you commute to every day.
