Most people think of leadership as an occupation or a person who is formally in charge of others, but leadership is really the mechanism that enables a group to perform better. Specifically, leadership is a process of influence that enables a group of people to function as a team to achieve more than an individual or a badly led group. Leadership, then, is a resource for the group, and the critical issue is not what the leaders look like but how they influence the group.
The good news for those hoping to automate leadership is that its scientific study is well-established. Indeed, 100 years of academic research have enabled us to identify the key ingredients of leadership, so it is now possible to predict with a relatively high degree of accuracy whether someone will become a leader and how effectively they will lead if they get there. And once we are able to decode a phenomenon to break it down into its core components, then it is feasible to automate it. As Norbert Wiener, the father of cybernetics and a pioneer in robotics, noted: “If we can do anything in a clear and intelligible way, we can do it by machine.”
Unlike human leaders, a well-programmed robot would be selflessly focused on advancing the interest of its team
For example, a crucial component of effective leadership is technical expertise. Unsurprisingly, leaders make better decisions than their subordinates when they have higher levels of domain-specific knowledge and sometimes higher general intelligence than them. To the degree that this knowledge can be reduced to a fixed set of rules and facts, it would be hard for even the most experienced leader to compete with a machine.
Furthermore, while the logical and reasoning capabilities of humans tend to peak by the age of 30, intelligent machines can continue to learn and get smarter and faster as they process more data. Of course, a robot leader will not be able to replicate human intuition, but there is no real evidence that intuition – feelings about facts – makes leaders more effective. On the contrary, when intuition is not grounded on data it can produce toxic ideas and undesirable behaviors, such as prejudice, unconscious bias and discrimination.
Another key component of effective leadership is integrity, which involves putting the team ahead of the leader and displaying consistency between one’s words and actions. There are two main reasons for the importance of integrity in leadership. First, integrity is linked to trustworthiness and unless groups trust their leaders they will not be able (or willing) to perform well. Second, when leaders lack integrity they could engage in a range of unethical and counterproductive behaviours that harm their teams.
Given the frequency with which these toxic and destructive behaviours are displayed in leaders, including highly qualified and talented individuals at the top of successful and global organisations, it appears that the honesty bar is fairly low, so it should not be difficult to design robot leaders that outperform most of their human counterparts on this score.
Needless to say, unlike human leaders, a well-programmed robot would be selflessly focused on advancing the interest of its team – that would be its only agenda. In contrast, even when people lead effectively they tend to be driven by selfish and narcissistic desires (eg the need for status, recognition and power), which explains why they often derail. Indeed, one study estimates that up to 67% of managers can be expected to fail.
A third critical element for effective leadership is strategic self-awareness or the capacity to understand how one impacts on others. Self-aware leaders are able to examine themselves from other people’s perspective. They are alert to feedback and able the gauge how their acts and intentions may be interpreted by others, which enables them to proactively manage their reputation.
Although self-awareness might appear to be a human characteristic, it can be modelled in robots. Indeed, most AI systems comprise a feedback loop that enables them to adjust their decisions on the basis of environmental inputs (eg thermostats, chatbots and wearables). Meanwhile the technologies for identifying human emotions from audiovisual content are advancing rapidly. And again, it is not that this ability is particularly refined in leaders, which is why billions of pounds are devoted each year to executive coaching designed to help leaders increase their self-awareness.
A final key ingredient for effective leadership concerns good people-skills, often referred to as emotional intelligence (EQ). Leaders with higher EQ are able to stay calm and composed, even in stressful circumstances. They can read other people like a book and are capable of predicting and influencing the behaviour of others.
Although affective computing – the creation of emotionally intelligent systems – is still in its infancy, it is important to note that robots do not need to be able to feel in order to act in an emotionally intelligent manner. In fact, contrary to what people think, even in humans high EQ is associated with lower rather than higher emotionality: it is about controlling one’s impulses and inhibiting strong emotions in order to act rationally and minimise emotional interference.
EQ scores range from very low – with key characteristics being neurotic, hotheaded and emotionally hypersensitive – to very high, phlegmatic, impassive and unexcited, so the real challenge would be to create robots with low rather than high EQ.
Though the idea of a computer-generated manager may seem far-fetched at the moment, robot leaders could start entering the working environment and begin to outperform bad (or even average) human leaders within the next few decades.
By Tomas Chamorro-Premuzic
-About the Author
Tomas Chamorro-Premuzic is professor of business psychology at University College London, visiting professor at Columbia University and the CEO of Hogan Assessment Systems. He is co-founder of metaprofiling.com and author of Confidence: The Surprising Truth About How Much You Need and How to Get It.