IBM Watson. Google’s driverless cars. Uncanny recommendations from Amazon, Siri, and Google Now.
Not to mention the Financial Times/McKinsey Business Book of the Year Award for 2015 going to Martin Ford’s Rise of the Robots: Technology and the threat of a jobless future, summarized as:
What are the jobs of the future? How many will there be? And who will have them? We might imagine—and hope—that today’s industrial revolution will unfold like the last: even as some jobs are eliminated, more will be created to deal with the new innovations of a new era. In Rise of the Robots, Silicon Valley entrepreneur Martin Ford argues that this is absolutely not the case. As technology continues to accelerate and machines begin taking care of themselves, fewer people will be necessary. Artificial intelligence is already well on its way to making “good jobs” obsolete: many paralegals, journalists, office workers, and even computer programmers are poised to be replaced by robots and smart software. As progress continues, blue and white collar jobs alike will evaporate, squeezing working- and middle-class families ever further. At the same time, households are under assault from exploding costs, especially from the two major industries—education and health care—that, so far, have not been transformed by information technology. The result could well be massive unemployment and inequality as well as the implosion of the consumer economy itself.
In Rise of the Robots, Ford details what machine intelligence and robotics can accomplish, and implores employers, scholars, and policy makers alike to face the implications. The past solutions to technological disruption, especially more training and education, aren’t going to work.
It all sounds dire enough, but today I want to step away from the dueling camps AI and sophisticated technology in general seem to call out in people (I’ve called them The Believers and The Deniers) and suggest a more practical approach for how you and your firm ought to address this right here right now.
As regular readers know, I often find McKinsey’s analyses to be refreshingly free of dross, cant, and embedded bias. There’s a reason they’re McKinsey, I suppose. So our text for today is Four fundamentals of workplace automation, which presents a useful framework for thinking about how AI is and may in future actually play out in workplaces across the country and across the globe. As we’ve come to find customary with McKinsey, the dataset underlying their analysis is exhaustive: In this case, breaking down actual workplace activities into roughly 2,000 individual procedures and then overlyaying on top of that “18 different capabilities that potentialy could be automated.”
To give you a flavor of what this actually means, here are some of the activities, using retail as an example:
- greet customers
- answer questions about products
- demonstrate products
- process transactions
- maintain the store itself.
And the capabilities:
- social and emotional sensing
- understanding natural language
- generating natural language
- recognizing patterns
- retrieving information
- being mobile
- being able to navigate a physical environment.
Can you say “granular?” That of course is the point. And it leads them directly to their core insight: It makes no sense to talk or think in terms of automating “jobs,” “occupations,” or, more pointedly for the current audience, “professions.” It only makes sense to talk or think in terms of automating “activities”—which require underlying “capabilities” if they are to be performed satisfactorily. By the way, if you’re wondering about computers’ ability to understand natural language, this past summer Apple claimed that Siri was down to a 5% word error rate, which is far from perfect but realistically good enough for most day to day interactions. More importantly to my way of thinking, Apple also reported that Siri was serving over a billion requests per week, which reinforces the power of relentless continuous improvement that’s part and parcel of technology these days.
So if we recast the debate in those terms, where do we stand?
According to our analysis, fewer than 5 percent of occupations can be entirely automated using current technology. However, about 60 percent of occupations could have 30 percent or more of their constituent activities automated. In other words, automation is likely to change the vast majority of occupations—at least to some degree—which will necessitate significant job redefinition and a transformation of business processes.
Right around here many of you—this is just a hunch, folks—are probably imagining that that “30% of the 60%” estimate is highly correlated with relatively low-wage, low-skill jobs. Our authors would beg to differ, and they have an r2 calculation to prove it.
Specifically, they added into their data mix of (a) occupations; (b) work activities; and (c) capabilities, a fourth metric, namely (d) wages for each occupation. Now they can chart a comparison of wages and automation potential:
There is indeed a significant correlation (p-value <0.01) but the correlation is highly variable and wages alone fail to explain most of it (r2 = 0.19). This actually comports thoroughly with common sense. Some significant percentage of a CEO’s time (or yours) is spent digesting reports and playing air traffic controller with incoming requests, after all—neither exactly high-intensity intellectual or creative challenges—while much of what waiters, groundskeepers, nurses, and barbers do isn’t susceptible to automation on any plausible cost/benefit analysis.
So: What to make of all this?
I prefer to recast the cliche’d “The Devil is in the details” as “God is in the details.”
Trying to reach grand conclusions about will or won’t lawyers’ jobs be automated out of existence is the wrong question and leads to the feckless and circular dueling monologues of the Deniers and the Believers. Rather, you need to ask yourself, and pay intense attention to, “details” such as:
- the specific capabilities of automation to perform (highly specified) activities lawyers currently perform;
- the speed with which you and your competitors can adopt targeted technologies and redefine jobs that may have been formerly centered around now-automated activities;
- needless to say, the cost of all of this and the payback cycle on the investment, retraining, and re-engineering of your firm’s processes;
- and whether clients perceive more automation as enhancing efficiency and quality or compromising it. (I said “clients” very pointedly: This is not your call to make.)
Finally, the irrepressible optimist in me must close by revealing another finding of our McKinsey friends in their granular exploration of what workers actually do all day: What do you suppose they have to say about sensing emotion and sheer creativity, two capabilities that humans are, so I choose to believe, hard-wired for but at which machines are pathetic and may remain so for a long time into the future?
The bad news is they found:
Just 4 percent of the work activities across the US economy require creativity at a median human level of performance. Similarly, only 29 percent of work activities require a median human level of performance in sensing emotion.
I choose to interpret this not as a sad commentary on the dumbing-down of the American workplace (yes, of course it is that) but as a fat opportunity to introduce more demand for human interaction and creativity into all our daily lives. One step at a time.