Is the economic singularity near?Leave a Comment
How times have changed. Just a few years ago ideas like „technological unemployment“ or the „technological singularity“ made you either a crackpot or „that crazy future guy“ in the eyes of most people. For example, when I offered to write an article on the latter during my first year here at prisma, it was literally rejected due to „sounding like a conspiracy theory“. Now, these concepts are rapidly gaining social acceptance and are discussed in places as the Nobel Prize Dialogue, the World Economic Forum or even the UN and many reputable professors, CEOs and researchers in the fields of AI and economics have endorsed the Open Letter on the Digital Economy, the Open Letter on Autonomous Weapons as well as the Open Letter on Artificial Intelligence.
Of course, the propagation of a meme does not imply its validity. Indeed, those more skeptical about the pace of technological innovation will rightfully point out that, at least in terms of technological unemployment, we have already been here more than once. The Luddite riots in England and the correlating „Maschinensturm“ in German speaking countries during the First Industrial Revolution being the most prominent example. Technological unemployment has been an issue ever since the invention of the wheel and all the troubles of restructuring as a temporary increase in unemployment aside, this has been a good thing. 95% of Swiss people are essentially technologically unemployed farmers and most of them seem pretty happy about it. So, why should this time be different?
Smart people have coined different terms for the age we are entering. Brynjolffson & McAfee, whose book was recently recommended by Thomas Bieger, call it the „Second Machine Age“, Jeremy Rifkin, whose ideas influenced China’s latest 5-year-plan, calls it the „Third Industrial Revolution“, WEF-Founder Klaus Schwab tops that and refers to it as the „Fourth Industrial Revolution“, as does the CEO of automation & energy giant ABB. However, in order to be fundamentally different from previous revolutions, whichever way you count them, we would have to reach what some call the “economic singularity”, a point at which automation sustainably destroys more jobs than it can create.
People say robots and mean AI
The recurring theme of robots stealing your job, or taking over the world for that matter, is a human fantasy largely unfounded in reality. I plead guilty for having indulged in that trope myself back in 2014 when I wrote “Roboter sind die neuen Ausländer”. At least, I can claim to be in good company since some of the best books on technological unemployment as Federico Pistonos “Robots Will Steal Your Job, But That’s OK” and Martin Fords “The Rise of the Robots” do focus on robots in their titles as well. Also, from an etymological point of view it really does make sense to talk about robots and the future of work, since the word “robot” literally means worker.
However, talking about robots is misleading because the main threat of technological unemployment (or taking over the world) doesn’t come from the field of robotics and some of the most impressive progress in robotics is actually spillover from advances in machine learning. When we think of robots, we usually picture some humanoid electro-mechanical machine, which is probably also the reason for the popularity of the robot theme. We humans are intelligent and therefore many seem to think that the more human-like its appearance, the more intelligent an artificial agent has to be. People are impressed and creeped out by the most anthropomorphized robots as for example “Sophia”, which was presented at this years SXSW. However, if you think about these “uncanny valley” robots, they are essentially chatbots dressed up with silicone masks and the robotics part of it really isn’t that impressive.
If anything, robotics is lagging behind in terms of automation. The “Technology at work 2.0” report from Oxford University and Citi Bank actually lists “manual dexterity“, the ability to quickly move your hand, your hand together with your arm, or your two hands to grasp, manipulate, or assemble objects and „gross body coordination“, the ability to coordinate the movement of your arms, legs, and torso together when the whole body is in motion, as two of the least automatable skills. So, for example, while we can already see the ordering process being automated with in-house tablets and increasingly your smartphone, serving food will mostly remain a mostly human task, at least in the near-term future.
Intelligence isn’t magic
The fundamental difference between a manual car and an autonomous car isn’t the coachwork; it’s in its algorithms. In it’s essence the next wave of automation is about intelligence.
“Now comes the second machine age. Computers and other digital advances are doing for mental power—the ability to use our brains to understand and shape our environments—what the steam engine and its descendants did for muscle power.”― Erik Brynjolfsson
In the short-term this means domain-specific narrow artificial intelligence and if the history of artificial intelligence is any guide, since it works, we won’t call it AI, but use “less threatening” terms like chatbots or machine learning instead. Artificial narrow intelligence as well as a number of future and emerging technologies as robotics, drones, augmented reality, virtual reality, 3D-printing, graphene, blockchain, big data or the Internet of Things can and will have a profound impact on the way we live in the future. However, the real game-changer is artificial general intelligence. I am not saying the technologies above may not lead to technological unemployment, social unrest or a restructuring of the economy, nevertheless, the economy would still remain largely human-driven (in a non-literal sense). While different tasks in different countries reach their “peak jobs” at different points in time, the global “peak jobs” will probably occur relatively close to the advent of artificial general intelligence. The nearer we get there, the harder it will become to allocate all human labor supply in types of tasks at which humans are still competitive.
Greater than human artificial general intelligence may seem like a pipe dream to many right now, but the truth is that we aren’t magical and our brains aren’t magical. We don’t test our medications on mice, because we greatly care about their health, it’s the exact opposite actually. Or to put it in other words:
“Far from being the smartest possible biological species, we are probably better thought of as the stupidest possible biological species capable of starting a technological civilization – a niche we filled because we got there first, not because we are in any sense optimally adapted to it.” –Nick Bostrom
We don’t live in a state of equilibrium. Once we had passed a certain threshold, we humans started a positive feedback-loop and technological progress won’t magically stop at any time soon. So, absent of any catastrophe, we will reach the “economic singularity” and, probably a little after, the technological singularity, however, the details, as for example, whether we will see an ever sharper digital divide or whether humans will survive at all, are still up for discussion.
Don’t wait for the future to happen, shape it!
So, how long until your (future) job can be automated? Will we see sudden spikes in unemployment due to technologies like self-driving cars? Do we already need to restructure our economy and social security system? What about a universal basic income? What are the ethical conundrums arising from these new technologies (trolleyology, attribution, employment, artificial consciousness)? How far away are we from human-level artificial intelligence? How big is the existential risk from AI?
Come and join the panel discussion “The Rise of the Robots: Will they take our jobs?” on Monday, 25.4, 18:15 in Room 09-011. The panelists are Prof. Dr. Jana Koehler, Professor of Computer Science at the University of Lucerne, David Iselin, Researcher at the KOF Swiss Economic Institute at the ETH, Adriano Mannino, Philosopher and President of the Effective Altruism Foundation and the computer scientist and co-founder of Syntacts, Kaspar Etter, who will also be the moderator. The event is organized by Effective Altruism HSG and afterwards there will be an aperitif giving you the opportunity to discuss the topic with peers and panelists.