The leadership dilemma of AI…or…Leadership, the Next Generation.

by Sarah Firisen

Screen-Shot-2017-08-15-at-7.48.21-AMWatch this video. No, I mean, right now, go and watch this video, I’ll wait. Even if you don’t agree with all of it, even if you think it’s unnecessary scaremongering, you should still find it thought provoking and at least a little scary (most people find it terrifying), and if you don't, then you’re really not paying attention. If you can’t be bothered to watch it, the basic premise is that we are very very quickly, far quicker than most of us realize, moving towards a world with so much automation that “Humans need not apply” for most jobs. That we are moving towards a very near-term future where humans, like horses in the past, aren’t just unemployed, they’re unemployable.

Bill Gates famously wrote, “We always overestimate the change that will occur in the next two years and underestimate the change that will occur in the next ten. Don't let yourself be lulled into inaction." It might be tempting to brush off predictions of a future without human work. And indeed, today, the likelihood of such a future does depends who you ask; the US government says not to worry (and this was before the current science denying administration), the UK government is less convinced.

By some predictions, more than half the human race could be unemployed, and more importantly unemployable, by as early as 2045. Uber recently bought Otto, an autonomous truck company. In October 2016 they made their first delivery, 50,000 beers. In the US alone there are 3.5 million truck drivers. That’s a lot of jobs and a lot of people to find additional employment for, but it’s not half the human race by any standards – even though transportation is the largest employment category in the world. Nevertheless, maybe it’s just, as the UK Science and Technology Committee says, that “Human beings must develop new skills to compete in a world where artificial intelligence is becoming more prevalent”. This is hardly a new problem; the 19th century Luddites were a group of English workers who destroyed machinery, particularly in the mills, because they feared it would take their livelihood away from them. They were right, it did, but those weren’t great jobs. Mill work was strenuous, poorly paid labor that often led to chronic, sometimes fatal illnesses. The industrial revolution eventually significantly increased the standard of living for the general population.

The prediction of Jason Furman, chairman of the White House Council of Economic Advisers is that "AI will grow the economy instead of take jobs away. While some jobs may disappear, AI will create new jobs and consumer demand for new products and services”. But while there may be some truth to that, this isn’t like the Victorian industrial revolution, this is different, both in scale and in the kinds of jobs that are already being automated.

One big difference is that now it’s not just manual jobs at the lower end of the education scale that are being automated. Today, artificial intelligence, and more specifically machine learning, have advanced to the point where robots are increasingly doing more traditional white collar jobs. And they’re not just doing them as well as people, increasingly, they’re doing them better: IBM’s Watson recently diagnosed a rare form of Leukaemia in a patient that doctors were unable to diagnose. Honestly, the list of blue collar and professional jobs that can now be automated to one extent or another is so long that I could spend this entire article just listing them and still not be done. Suffice it to say, whatever job you do, odds are significant portions of it are at risk for automation of some kind.

I’m not a politician, or an economist, or a sociologist. I’m not going to talk about what a huge problem this is likely to be for the human race or even attempt to put forward a solution. I’m going to look at a very small sliver of the potential issue: development of leadership skills. Many careers, to one extent or another rely on the master apprentice model. But if the apprentice’s tasks are automated, how will anyone ever develop into the master? And this is only one side of this challenge, partners in law firms don’t just magically happen. As in most professions, the process is one of winnowing the field by watching young people do the junior roles, learn and grow. Based on this, assessments are made about who has what it takes to move into leadership roles.

As the video above says, legal discovery is a perfect robot job – people get tired, make mistakes, get distracted. Computers can analyze thousands of pages of documents with a level of precision and stamina that people are incapable of matching. In theory, associates can be freed up to perform higher level work. But as automation becomes more pervasive, it’s hard not to conclude that will actually happen is that law firms need less associates. It's obvious what the incentive is to the clients, the partners and the law firm overall; fewer salaries, less mistakes, more wins.

In Star Trek The Next Generation, Mr. Data is an extremely human-like android who is able to plug into computer systems, crunch enormous amounts of data and work alongside the humanoid members of the crew seamlessly. He can’t get sick, any injuries can be fixed mechanically, he doesn’t need to eat or sleep and despite his lack of an emotion chip, clearly is able to be “moral”, build relationships and generally function as a pretty senior member of the crew. So given that he does everything the non-androids do and more, why hasn’t the United Federation of Planets worked to make most or all of its crew’s androids like Data? And why are the people who work alongside of him not more worried that this will happen one day?

Maybe you don’t want the ship to be led by an android, I’m not entirely sure why not, but let’s say that there are elements of leadership that need a human touch. So in the world of Star Trek, you have the captain and maybe the second in command and maybe a few others who have to be human. How do you develop those people? If you don’t have Wesley Crusher sitting there as an ensign, watching Picard leading, learning from his own experiences and those of the crew around him, how do you end up with a Captain Crusher one day?

This gets to at least a part of the dilemma posed by machines in the workplace: with those lower level skills being automated, how do you train and develop people to have the higher level skills one day? There’s a longer term leadership development issue at play as this extreme workforce transformation encroaches into more and more professions: today, a law partner is able to represent their client, advising them, arguing for them in court, because they came up through the ranks as a legal associate and learned the craft. And if you’re in leadership, if you don’t get have benefit of decades of winnowing the field watching young people do the junior roles, learn and grow, how do you know who has what it takes to become the next Captain Picard?

I met a young ER resident recently and asked him if he, someone starting out in medicine, was worried about increasing automation in the field. He totally pooh-poohed the whole idea. He said that even if aspects of medical practice can be automated, computers will never be able to do what doctors do. I was skeptical and said “but if what doctors do is apply a set of scientific rules to data, doesn’t that make them actually perfect candidates for automation?” After all, no matter how much experience a doctor has, how many cases of something they’ve seen over the years, they can never have access to the amount of case data that a computer can. They can never know all possible drugs and drug interactions – but a computer could. His answer was that a lot of what a doctor does comes from the gut rather than being the data and rules-based scenario I described. He may be right but it’s pretty terrifying if he is, because after all, as we all know, doctors are not infallible. I think I’d rather my doctor based his diagnosis on data and science and not his or her gut. But again, let’s say that we’re both right; that there is a lot of what doctors do that is perfect for automation at the lower end of the patient spectrum. But that there is a degree of instinct that comes from having seen it all over a long career, from understanding human nature etc, that won’t be automated anytime soon. Again, like the legal example, if we don’t need young doctors on the wards monitoring patients and making routine care decisions, how do they develop?

Of course, you could look at the other side of this; there are lots of professions that used to involve activities that were considered foundational to growing professionally. Computer programming used to involve understanding machine code, understanding memory management etc. Most of today’s programming languages are of a higher level without such requirements. It doesn’t mean that today’s programmers aren’t as good it just means that they’re able to spend their time doing higher level, more sophisticated things. It’s a natural part of human evolution that lower level tasks become automated allowing us to focus on higher value tasks. We’re clearly moving to a place where being a good speller is irrelevant, most people don’t hand write things these days and we have spellcheck. Does this really matter? I’m sure to some people it seems awful. But if children can use time they used to spend learning to spell learning higher level skills that are more valuable in today’s world, is it so awful? And so perhaps apprentice-like jobs will evolve in the same way. In many ways they already are and have been for a long time.

I think the big difference is speed and scale. This unprecedented automation of jobs is moving at an exponentially fast pace. The normal evolution of job skills is being disrupted and it’s unclear that we’re even fully aware of this, let alone ready.

In the Star Trek future, hunger and poverty have been eliminated.Humanity has dedicated itself to exploration and self-improvement. And indeed, one of the suggestions that is being looked at for the answer of large scale human unemployability is the Universal Basic Income, something that is being tested with mixed results in Finland. But if everyone’s basic needs are taken care of, how do you motivate anyone to become a scientist or to sign up to man your starships for exploration? Even if we eventually get to a point where humans are totally unnecessary for any of these roles, there will inevitably be a transitional time where, like a Starship captain, human leadership is still needed. In a world where humans need not apply for most jobs, how do we make sure people still have the skills needed to captain the ship?