Ever since The Guardian published Robot revolution: rise of ‘thinking’ machines could exacerbate inequality, I’ve been waiting for the report. My hope? To get beyond the intellectual stalemate in the robots-versus-jobs debate.
Well, as The Guardian noted in its follow up, Artificial intelligence: ‘Homo sapiens will be split into a handful of gods and the rest of us’, that report is now publicly available (see PDF link).
The Bank of America/Merrill Lynch (BAML) report, entitled Creative Disruption, is an “equity strategy” report for “thematic investors”. The Guardian claims it’s a 300-page report – I count 27 pages but no matter, maybe I got the easy reader version.
Judging from The Guardian’s sensational headlines, you’d think the report was an apocalyptic view of robots consuming human livelihoods with impunity, while income disparity widens. But it’s actually a measured view that contrasts “bull” and “bear” takes on our economic futures.
The report is a mix of informative charts and “the consumer wins” digital cliches. Some predictions:
- IoT is expected to be a $7tn industry by 2020
- The Sharing Economy has a potential market opportunity of over $450bn
- Local consumer On-Line Services has a potential market opportunity of $500bn
Eye-watering, but also useless. Anyone can breathlessly float “potential” markets. More illuminating is the contrast of bearish vs. bullish views on robotics and productivity.
The “Bear Case” on big technology
BAML presents the bear case as:
- big technology gains in growth, income and productivity are largely behind us
- today’s technology is more “disruptive” than “creative” for economic activity
The bear case is supported by Economist Robert J. Gordon. In a National Bureau of Economic Research (NBER) paper, Gordon argued that big tech advances like the internal-combustion engine, telecommunications and electricity boosted productivity in documented ways. The computing revolution doesn’t have the same productivity impact, as per the “productivity paradox,” the apparent contradiction between advances in computer power and comparatively slow economic growth.
Another bearish take comes by way of Peter Thiel, a venture capitalist who notes sluggish progress in transportation and medicine since the early 1970s. Thiel is the curmudgeon who quipped, “We wanted flying cars, instead we got 140 characters.” (I wonder if Thiel will change his views now that he can send a longer direct message).
The bear case rounds out with “technological unemployment” – and this is where the robots come in. The idea is that automation/robotics inhibits job growth, which in turn stunts consumer spending. But as BAML points out, the unemployment rate in the US is just 5.5 percent (though that does not take into account the “permanently unemployed” who are no longer seeking work).
Supporting a grim view of robots are stats like these:
- The number of industrial robots is up 72% in the last ten years
- The number of US manufacturing jobs is down 16%
- Oxford University recently predicted that 45 percent of all US jobs will be replaced by robots within the next twenty years
- In the future, BOML sees machines increasing their task load in banking, logistics, health care and other service sector industries, displacing pesky humans
The big tech “Bull Case”
The Bull Case also has some significant stats. Typically, there is an extended time lag between innovation and the materialization of a technology. BAML cites:
- It took decades after the innovation of electricity in the 1880s for economic growth attributable to electricity to take hold
- Economist Chad Syverson’s research shows that productivity gains from tech are proceeding at a similar pace to electricity
Adding to the bull case, BAML cites the book The Second Machine Age by Erik Brynjolfsson and Andrew McAfee:
[The authors argue] that the global economy is on the cusp of a dramatic growth spurt driven by smart machines that can take full advantage of advances in computer processing, artificial intelligence, networked communication and the digitization of just about everything.
They suggest that we are on the cusp of seismic change fueled by accumulation of advances such as the capacity of algorithms to handle data growing exponentially, autonomous vehicles, three-dimensional printing, and so on.
Adding to the optimism is Ray Kurzweil, director of engineering at Google, inventor, and futurist. Kurzweil argues that the period for technological innovation to become mass-market goes down at each phase in innovation. For example, Kurzweil predicts that by the time we get to 2023, a $1,000 laptop will be able to calculate at the speed of the human brain. The result of this evolution will be the so-called “singularity,” when AI exceeds the abilities of human intelligence.
Futuristic stuff perhaps – but as The Guardian notes, it ties into a prediction made by economist John Maynard Keynes, way back in 1930. Keynes predicted that within a century, technology will reduce the working week to 15 hours – with the rest of the time devoted to leisure (I don’t think we are trending in the right direction on this one).
The BAML report doesn’t take a final stand on the bear-versus-bull debate, opting to drill into issues such as cybersecurity, deflation, and energy-producing “Cleantech” instead.
The news headlines imply a grim outlook, but that is not the tone of the report. The Guardian itself quotes one of the authors as saying, “It’s not meant to be a doom and gloom report: one of the ways we think people could help themselves here is through education.”
The deeper report sections are worth a look, but do not move us past the robots-versus-jobs stalemate. I thought the bullish views were weakly presented against the bearish, but that doesn’t settle the matter. (the Cleantech section on renewable energy carries way more legitimate reasons for optimism than the hypothetical singularity).
Are we back to where we started? We’re still divided on whether robotics and automation will fuel growth and create more jobs than they displace. On another level, there is a conflict between those who view machines as freeing up humans from demeaning work, and those who believe that automation fuels skills/income disparity, dividing humans into digital overlords or paid-by-project “task rabbits” – with little middle ground. That utopian/dystopian divide is well beyond my scope.
But there are three questions worth focusing on:
- We all agree automation will displace some workers (e.g. taxi drivers) – even if that “disruption” can be rationalized as inevitable and not ultimately harmful. What kind of reskilling/education needs to be implemented for the most impacted industries?
- To what extent will robotics move “up” into tasks that require greater analytical capabilities and intellectual discernment? Can we accurately chart the pace of that upsurge and plan an equivalent reskilling?
- How does our educational system need to overhaul so we can give students a chance to succeed rather than bid against each other for the privilege of designing self-driving cars?
At the risk of ruining my own editorial calendar, there is enough consensus to pause the debate and bear down on these problems.
To be continued: note that our Chris Middleton has a piece coming Thursday on the impact of robotics and banking, including BAML’s own plans for robotics and how this ties to their new report. That piece is now live: The robot banking machine doesn’t add up.
Image credit: Robot observando una rosa © anibal – Fotolia.com