Are Robots Going To Replace CEOs?

A few weeks ago, I got into an intense discussion with a good friend, who’s a surgeon. He does important work, literally saving lives for a living. But I said his days in the operating room might just be numbered: the robots are coming. Already, driverless trucks are making their way across Europe, and “virtual employees” are starting to replace call center jobs. By some counts, surgery is not far off on the machine-learning hit list. Indeed, Google’s DeepMind artificial intelligence group, which just mastered the ancient, inscrutable game of Go, is now tackling eye disease. It’ll be studying the record of 1.6 million patients (!) in London hospitals to learn how to best treat cases.

But then my friend threw the same possibility back at me. Well, what makes being a CEO so special? How do you know the robots aren’t coming for your job? This got me thinking. I lead a half-dozen online optical companies with almost a thousand employees around the world. But what are the possibilities of my role being replaceable? Is there a risk that CEOs will be made redundant too, when the robots come?

The promise (and limits) of AI

In his new book The Fourth Industrial Revolution, engineer and World Economic Forum founder Klaus Schwab explains that one of the ‘mega-trends’ reshaping society - like the automobile and the smartphone before it - is the wide application of artificial intelligence. An obvious, but key distinction, needs to be made here between automation and AI. Automation is just a machine performing a rote, repetitive task - like the work robots currently do on auto assembly lines. With AI, machines actually observe and learn new capabilities - on their own.

That’s the essence of machine learning. A programmer doesn’t have to write every line of code for an algorithm to execute a function; you just have to design the algorithm to learn from the “experience” of a data set. It’s not like Google’s developers wrote every line of code for self-driving cars; rather, the algorithms observed real life drivers and now draw from that database to make the right decision.  

But right now, machine learning still faces serious limits. One of the best thinkers on artificial intelligence is Brown University computer scientist Michael Littman. As Littman describes it, algorithms thrive in environments where there’s very fixed, very static rule sets for what success or failure look like. Board games have fixed, rigid rules, which is why computers have been so readily able to tackle them. But other fields are much more dynamic, the rule sets much more fluid. Robots aren’t going to master figure skating anytime soon, for example. Artistry, expression and even intuition are hard to build into an algorithm … at least for the moment.

The downsizeable CEO?

So where does that leave CEOs? Well, that depends whether you look at the role as a collection of predictable, repeatable tasks - planning budgets, evaluating talent, assessing competitive threats - or as more of an art: something that requires the kind of discernment, intuition, emotion and flexibility that’s beyond the reach of an algorithm.

I’d like to think that my day is filled with exactly those kinds of irreplaceably human tasks. But that’s not necessarily the case. In fact, lots of responsibilities seem well suited to machine intelligence. Especially for early stage companies, for example, hiring is of paramount importance. The first dozen or so people you bring on board are going to set the trajectory for your culture. But an algorithm, armed with the right data about how employees fit into organizations, might assemble a much more cohesive team than I ever could: sorting through the usual materials like resumes, references and salary requirements, but also factoring in far more nuanced and complex variables like personality cues and interpersonal matches.

The same goes for other functions at the core of the CEO role, like business planning and analysis. This kind of quantitative aptitude is something that CEOs have traditionally needed to have in their quiver, but algorithms are already taking the responsibilities of Wall Street analysts. Billion-dollar decisions are being executed daily by intelligent machines. Certainly, an algorithm could one day do some of the heavy lifting for my quarterly budget.

A conductor for new technology

But there’s more to leading a company than numbers … and business obeys anything but fixed, static rules. Indeed, I think the place where CEOs are the most inimitable, the least replaceable is when we’re like those figure skaters - called upon for artistry, grace and improvisation. And, especially in an era of accelerating innovation, there’s plenty of skilled skating required.

Most critically, someone needs to be a “conductor” of all of these emerging technologies that are transforming business: selecting the right tools and ensuring that they work together to push the company in the right direction. This is already a key role for CEOs, and it will only get more important in the future - and, on a fundamental level, this function can’t be replaced by technology.

The canny CEO has a thorough understanding of the lay of the land; she knows what precisely is happening in her field - and adjacent fields - and incorporates the technologies that will inform her business, while abandoning the ones that will hold her company back. Mark Zuckerberg is a prime example: as the Internet and smartphones matured, Facebook didn’t just stick to social networking, it expanded into chat, and now Facebook Messenger is the most popular iOS app ever - after Facebook. And by buying up runaway hits Instagram and Whatsapp, and then blue-chip VR startup Oculus, Zuckerberg has shown that he doesn’t just understand where technology is at, but how to position his company for where it’s going. As CEO, he’s conducting divergent technologies like he’s standing on stage at Carnegie Hall, and it’s a model we can all follow.  

To be clear, I’m not suggesting that the future CEO needs to know how to code to be effective. Nor is this process of screening, selecting and implementing technology the only distinctly human thing that CEOs do. In fact, enumerating all the creative, unrepeatable and unforeseeable functions of the job would be impossible. At root, a CEO exists precisely to handle and respond to the unplanned and unexpected - most anything else can and should be delegated. For that reason, his portfolio of responsibilities is forever fluid and forever expanding.

Ultimately, the real work of the CEO is to be capable of responding to continuous change: to absorb the radically new and figure out how to act on that. With humble deference to my surgeon friend, the human anatomy today is much the same as it has been for thousands of years and much the same as it will be for thousands more. But I may wake up tomorrow and discover that a new invention has upended the optical industry, throwing the old rules out the window. That’s why a robot probably isn’t ready for my job - not quite yet.

This article was originally published in Fortune Insiders. What are your thoughts on whether robots could do the work of CEOs? Follow me for similar articles.