Here’s an educated guess: Among all of the possible things employees are likely to miss from their jobs since they have been asked to stop coming in to the office, their boss is probably not one of them. Why?
Because there is an odd tendency across organizations to promote people into leadership roles regardless of their talent, and without giving them much prior training on how to be a good boss. The flawed assumption here is that as soon as employees perform well in their role, or are at least able to persuade others that they have, they should be asked to do something else. Ironically, their reward for being a valuable asset to the organization is to be put in charge of others, have them manage a team, and in effect stop them from doing what they are good at: their job. The common result is that they then become a less valuable asset to the organization.
The perverse and irrational nature of this process was eloquently captured by the famous Peter Principle, which predicts that people will usually get promoted until they reach their own level of incompetence, suggesting that employees stagnate when they have reached a role that is one step above the one they should have stayed in. If you are no longer worthy of promotion, you should not have been promoted to your current role. An important extension of this principle, albeit focused on interests rather than skills, was recently proposed by the work column of The Economist. Thus Bartleby’s promotion curse states that people tend to get promoted until their job finally makes them miserable. Needless to say, this often includes making other people miserable, too.
Although smart HR professionals have been proactively pursuing a better way to appoint leaders, which is why they engage in elaborate philosophical discussions about competency taxonomies and high-potential models, the Peter and Bartleby principles represent a more honest and accurate depiction of how promotions are actually determined in most organizations, across jobs, industries and sectors, whether they like to admit it or not. Alas, past performance eclipses actual potential, and promotions are positively correlated with the number of direct reports added to your role, so you’d better learn how to deal with people, or at least not to hate it.
Unsurprisingly, most bosses are not known for their ability to manage. This was true before the pandemic, and will remain true during and after the pandemic, too. As if putting people with limited interest and talent for leadership in charge of others weren’t questionable enough, today bosses are faced with the additional challenge of learning to manage in a virtual-only environment, whether it’s Zoom, Microsoft Teams, or Google Hangouts, on top of having to deal with the chaos and disruption the pandemic has brought to their own their employees’ lives.
As one would expect, many people are struggling to define the blurred boundaries between what’s work and what’s personal, and back-t0-back virtual meetings are allegedly causing Zoom fatigue, turning us into Zoombies, and creating a state of technologically induced existential confusion where distinguish between the real, virtual, and imaginary is far from easy. The rules and guidelines of management, as well as our basic work etiquette, are being rewritten, such that even experienced bosses may fail to manage their teams effectively.
Perhaps technology can help?
After all, we are already dependent on it, so why not leverage as much of it as we can to perform our management duties more efficiently and productively? Furthermore, since the baseline for competent management is already pretty low, it would not take much for technology or any new tool to add significant improvements to employees’ work experience and performance, just by making their bosses a little more competent. Human relationships are inherently ambivalent, and our relationship to technology is no exception. We both love and hate it, like to complain about it, but cannot live without it: most people would probably have more trouble giving up WiFi than extending the lockdown and social distancing. Like the fetischist who fell in love with the foot but had to settle for the whole person, we are too deep into our relationship with technology to cherry pick only the parts we like, and it’s hard to have it both ways.
Any software or app that enables us to work remotely and communicate virtually, while retaining a large proportion of the features that make up analogue or real-world communication, is not just capturing an enormous amount of data on our work behaviors and relationships, but also learning to predict and understand ourselves, bosses, and colleagues, often better than they do themselves. As Yuval Harari noted, we have been hacked by AI and algorithms.
Outside of work, we seem pretty happy to be managed by algorithms, which is one of the reasons we have agree to dilute our privacy and share a great deal of personal data with them. We trust Amazon to tell us what we need (or at least want) to buy, Netflix to tell us what we should watch (to the point of gifting 5-hours precious hours of our life to Joe Exotic), Spotify to pick our music, and Twitter to curate our filter bubble, so we can stay partially informed and pleasantly uninformed. Imagine a future in which all these data are stitched together, and combined with our Whatsapp, Facebook, Uber, Oura, and browsing data, including our Google and YouTube searches, and you would get a pretty accurate and comprehensive picture of who you are. The more time we spend online, which increases when our offline activities are restricted or eliminated, the more integral our digital selves are to our personality, and there was never really a big difference between our online and offline personas anyway.
What about work, and specifically management? Could we invite algorithms to help our bosses with the basic tasks we expect every good manager to accomplish?
1) Measuring and monitoring our work-behaviors objectively, without conscious or unconscious biases.
2) Understanding what we do and why we do it, while ignoring irrelevant factors (not associated with our work, and not contributing to our performance)
3) Being consistent in their evaluations, without getting distracted, and leaving their own agendas aside to focus on what’s good for the organization.
4) Providing us with constant, helpful, critical feedback that enables us to improve our performance. This feedback should be data-driven, factual, and evidence-based, as well as include measurable indicators to monitor progress.
5) Our performance appraisal, including compensation and benefits, should be directly tied to the above metrics, such that two informed evaluators looking at our profile would reach exactly the same conclusion about how well we performed.
Competent managers do all of this pretty well, plus they add an indispensable human side – not replicable by machines – that includes empathy and caring (by definition, computers don’t give a damn). The problem is that they are a rare species, and even when you find them they are likely to do their job even better if they can harness and interpret granular behavioral data on their employees’ performance. This is where AI can help.
The main problem is that even when AI and algorithms could make an important contribution to elevating the average performance of managers, and in the absence of legal constrains for doing this according to the strictest ethical guidelines, most employees would still find this approach intrusive, untrustworthy, and rather creepy.
Just mentioning the possibility that algorithms could scrape our Zoom meetings to identify key behavioral markers from our verbal and non-verbal communication to infer if we are engaged, excited, upset, confused, or stressed, seems to freak people out. Even the notion that AI could mine our e-mails to detect some of the key verbal indicators of productive or unproductive behaviors, job satisfaction, and wellbeing, appears quite shocking to the average employee.
And yet humans are influenced by exactly the same data, which they are legally and ethically allowed to inspect, except they are far less competent at interpreting it. Last, but not least, when it comes to biases it is really quite hard for any AI to compete with humans. The only way is for AI to be trained to predict human preferences, in which case it will not just replicate, but also augment them. That is perhaps the biggest value that AI and algorithms have: when we scrape and analyze vast amounts of human data, we can differentiate between the factors that contribute to people’s success within an organization (meaning, their managers liked them), and those that actually contribute to their job performance (even when managers don’t want to accept it).
In short, AI may seem creepy, but it’s rarely as creepy as a human boss, unless we have trained it to emulate him or her. And there may not be a better way to spot creepy bosses than by using AI (creepy or not).