One day, not too far from now, we’re going to experience the greatest unemployment crisis in human history. Whether that will be good or bad depends entirely on how we prepare for it.
What happens when technology advances so far ahead that machine intelligence can perform all human labor? It may sound nice at first – perhaps even utopian – but the implications are more frightening than you might think.
This topic popped up when we looked at jobs that robots should take over, including many occupations that involve sales, data research, and transportation. In response to that post, one of our readers, Dmitry, posed an interesting question:
A much more pressing question is: what will “replaced” humans do? [Aside from the] ever-dwindling “elite” groups of creators, servitors and owners, there’ll be no “logical need” for the rest of the jobless – thus “non-earning” – population to exist in the capitalist paradigm.
Great question! While nobody can say for sure, plenty of minds brighter than my own have explored what such a scenario might entail. Let’s see what’s in store for us.
The Rise of Human-Level Intelligence
We are on the cusp of a tech-based societal transformation that will be at least as big as that of the Industrial Revolution. This isn’t just machine-driven automation for tedious manual labor. We’re talking about man-made creations that can “think” at a human level – or even beyond.
To be fair, that second-wave revolution won’t actually be here for quite some time. Optimistic projections put that kind of breakthrough in the 2040s or 2050s while more conservative models predict the 2080s or 2090s. So, we have anywhere from a few decades to close to a century.
But when you look at the advancements made in the past few years, it’s hard not to get excited over the possibilities. A few decades might seem far away, but that time is going to zip on by and the breakthrough will arrive before you know it.
For example, driverless cars are on the horizon. It’s not as sexy an idea as, say, hoverboards or flying cars, but it’ll still have an enormous impact on the way we live our lives and conduct commerce. Driverless cars aren’t generally intelligent in the same way that people are, but driving is a big, complicated, subtle cognitive task which is quickly moving into the reach of robots, and that’s a sign of things to come.
It’s not even necessarily clear that traditionally intellectual jobs are safe. Even in the near future, machines like IBM’s Watson system that famously won on Jeopardy may take over data-heavy jobs like doctor and lawyer, thanks to their ability to consume and integrate far more information than humans. In fact, traditionally intellectual jobs may be among the first to go. Ironically, some of the tasks we think of requiring enormous intelligence (like those that depend on an enormous amount of domain knowledge) are proving much easier for machines than relatively basic tasks like cleaning a house or making a burger.
We also shouldn’t rule out robotic creativity because it now seems that robots can be creative. Cutting edge robots today can compose music, write news stories, and paint artwork. Here’s a song written by a piece of software called “Emily Howell.”
Again, we’re not at a point where robots can compete head-on with human imagination, but these are steps in that direction.
What about on the battlefield? Microsoft has already developed a line of robotic security guards that are used to maintain security on one of its campuses. It wouldn’t require much imagination to take it one step further: autonomous war machines that are smarter and deadlier than humans. DARPA contractor Boston Dynamics has developed powerful humanoid robots with potential military applications, such as their latest quadruped, SPOT:
We’re making our robots smarter and better, but maybe we should be wary about where that might take us.
Robot Intelligence, Human Employment
Imagine it’s a century from now and our continuous progress in artificial intelligence has resulted in robots with human-level intellectual capabilities. In this hypothetical, robots can equal humans in all intellectual, mathematical, engineering, and creative pursuits.
What do you think the work force looks like in this scenario?
When the Industrial Revolution hit in the mid-1700s, people freaked out – the term ‘luddite’ comes from the anti-automation movement of this period. A single machine could match the production output of a hundred humans, essentially putting those people out of their jobs. Thus, machines lead to unemployment, right?
Think of every possible job in the world as an individual bucket and think of the people employed with that job as a drop in that bucket. For example, a farmer bucket. Here comes a set of machines that can do everything that a farmer does more cheaply, thus eliminating the need for the “farmer” occupation. Essentially, that bucket disappears.
But these ex-farmers are now freed up to work in whatever other buckets are available, thus boosting the production levels of those buckets — the economy as a whole is able to grow as a result of the added value created by all the cheap robot farmers. The result? An overall improvement in global production.
For now, machines are only good enough to replace jobs that require unskilled rote work, which means that they don’t cause unemployment but rather reemployment into other fields that can’t be automated. This point is worth emphasizing: the current employment crisis was not caused by automation.
But if we’re talking about a future time period when robots are on equal footing with humans and aren’t limited to rote automation, the situation is different.
Robots are cheaper and more efficient than humans. Humans need nine months to gestate, eighteen years to mature, and several additional years to train in a particular field, whether that means healthcare, engineering, the arts, or whatever else. In contrast, robots can be mass produced and once one has been programmed or trained, that data can be instantly and infinitely duplicated.
In other words, when there’s a choice between a robot and human with equal potential, the robot is always the more efficient choice. And when robots are skilled enough to take over every single job bucket on the market, there won’t be any buckets left for humans. That’s when reemployment becomes unemployment.
The Implications of “No-Humans-Required”
At this point in our hypothetical scene, two things are true: 1) robots are sufficient to produce everything necessary for humanity’s continued existence and 2) the vast majority of humans can’t find work because all work is done by robots. What would this kind of society look like?
The first option is that nothing changes, and we find ourselves in a world in which a small fraction of people who were independently wealthy before the mass automation can live comfortably, reaping the benefits of the rapidly growing robot economy. The poor would either starve or be taken care of by charity from a small cabal of wealthy benefactors. This scenario can range from ‘pretty okay’ to ‘absolutely nightmarish,’ depending on the details. Incidentally, this is also the default option.
But let’s say we were to try to change the system to avoid this scenario. What might that look like?
The first thing to go would probably be market economies.
Markets are based on needs and wants. Money represents the exchange of one need or want for another. When all goods and services can be produced at perfect capacity by non-human labor, the utopian scenario would be perfect socialism: humans can reap the benefits of a vast robot economy without having to pay for them, and everyone’s needs and wants are fulfilled without the need for money.
This would free up humans to pursue whatever they want without any material limitations. Since nobody would be producing or buying or selling anything, all sense of ownership would quickly disappear – and if something were to break or be lost, that thing could be replaced almost immediately.
It wouldn’t spell the end of all human interactions, though. Money might cease to exist, but another kind of economy might take its place: an economy based on intangible goods such as status, reputation, skills, or creativity.
Yet no matter what happens, the global robot force would still succumb to entropy. Broken robots would still need to be repaired, lost robots would still need to be replaced, and materials and energy would still need to be gathered. Who’s going to be responsible for this?
One possibility: humans. These might be the only real jobs remaining in the world. While 99.9% of society lives free from work, a small segment would have to slave away.
In the worst case, a small caste of laborers would become slaves for life. In a not-as-bad case, society could enforce some kind of mandatory rotation or incentive system where skilled people are somehow encouraged or forced to take shifts working as robot maintainers.
All of this assumes that humans remain superior to robots, but that may actually be a foolish assumption to make.
If robots were to achieve a level of intelligence that matched the human mind, it’d be reasonable to assume that they’d also be intelligent enough to take care of themselves. Such an intelligence would imply the ability to self-replicate, self-repair, and even gather their own energy resources. Sounds useful, right?
But there’s something called an “intelligence explosion,” which describes an intelligence threshold that, if surpassed, would allow robots to accelerate their own intelligence beyond the intelligence of humans, resulting in a kind of robotic super-intelligence appearing suddenly and with little warning.
Are you smarter than an ant? Of course! But it’s a silly question. Human intelligence and ant intelligence are on such different levels that the question is meaningless. If robots became super-intelligent, the gap between them and us may be as large as – or even larger than – the gap between us and ants.
And at that point, we’d lose any illusion of control over our creations.
That’s not to say that robots would necessarily be hostile towards us, but they would be controlled by the goals we gave them. If those goals aren’t compatible with ours, things could get ugly very quickly. For example: if the robots need more fuel to do their jobs and they decide that the most efficient way to obtain more fuel is to break down humans and other organic matter into our raw hydrocarbons… well, that wouldn’t be a happy ending for us.
It’s hard to summarize this kind of speculative discussion because there are so many possibilities that are just too far in the future, but here’s the gist of it: while machines aren’t the cause of modern economic woes, they will have a profound effect within the next century, and we need to be prepared for it.
What will we do when the line between artificial intelligence and human intelligence disappears? The optimist in me wants to hope for a utopia, but the realist in me can easily imagine the end of humanity as we know it.
What do you think? Will robotic super-intelligence lead us towards salvation or poverty? Share your thoughts with us in the comments below!
Image Credits: you are fired Via Shutterstock, Robot hand Via Shutterstock, Robot Job Comic Via Shutterstock, Open Relaxation Via Shutterstock, Gear Arm Via Shutterstock, Intelligence Explosion Via Shutterstock, Robot Human Mind Via Shutterstock