In a recent series of tweets with Nolind Whachell and Perry Timms, building on the recent post here The way we work doesn’t work, anymore, Perry wondered:
To which I responded:
And then Nolind wondered:
And then I replied:
Driverless management is a concept I’ve been noodling for some time, one that I intend to explore in the next weeks and months. I introduced the idea in my predictions for 2017, but I have now returned to expand on the trend.
Driverless management takes off where self-organization and autonomous work leave off. In the immediate near-term, artificial intelligence will have evolved to the point that a great deal of what managers and others in organizations do (especially ‘human resources’) will be automated. The first inroads have already been made in areas where people do a terrible job, like hiring and promotion.
Consider technologies like Cornerstone, where large companies have basically outsourced hiring their call center staff to an algorithm: algorithmic HR. Google looked at its history of hiring — despite all the hoohah around its fabled brain teaser interview questions — ‘how many pingpong balls can fit in a school bus?’ — and discovered that almost no one was good at hiring. Which is to say that no one was consistently good at hiring people who were successful in their careers at Google. It was random, for all intents and purposes.
The number one reason people quit their jobs is bad bosses. Instead of trying to work around the problem by training bosses to be better — which doesn’t seem to help much — we should move as aggressively as we can to apply narrowly-focused AI to the various activities that we tie a strong around and call ‘management’.
The number one reason people quit their jobs is bad bosses.
Driverless management relies on AI and ‘sensors’ in the same way that driverless cars do.
The ‘sensors’ in this case will be accessing all sorts of information about the workplace and the people in it: communication patterns, social network analysis, people’s actions and reactions. Basically, we’ll start turning all the magic we’ve developed to understand people’s behavior as consumers and repurpose it to understand people’s behavior at work.
And the AI in driverless management will ‘steer’ the business through mediation of human interactions, and working to make sure that work culture principles are applied in decision-making, negotiation, and allocation of resources.
We’ll start turning all the magic we’ve developed to understand people’s behavior as consumers and repurpose it to understand people’s behavior at work.
This might be the fullest realization of servant leadership, since the AI involved would be an asset owned (or rented) by the business, without its own agenda or agency.
Yes, I anticipate a great number of management enthusiasts will come forward, making one or more of the following arguments:
- But great managers exist.
If so, they are few and far between, at the best. We have enormous levels of worker disengagement, questionable productivity, and grinding bureaucracy. If managers are so great, why is work so bad? Anyway, the existence of a few great managers doesn’t mean we should maintain a system of coordination that requires more great managers than we can produce. It’s doomed to failure. And please don’t tell me we just need another few decades, or better tools, or better training to get it right.
If managers are so great, why is work so bad?
2. But we need leaders to inspire us, to train us, to lead us.
Yes, people do need to be trained. Or may be better said, people do better when well-educated, and trained in the particulars of their work domain. However, increasingly that is undertaken prior to getting a job, and whether off- or on-the-job, training is increasingly the responsibility of the individual, perhaps supported by co-workers and managers.
Inspiration? Yes, but charismatic sociopaths are inspirational, and they have a tendency to lead people off cliffs. Personally, I think we’d be better off working bottom up on the inspiration angle, where each person would find meaning and purpose in their own relationship to their vocation, and not get it secondhand from some entrepreneur who believes he was born with the CEO gene (and I use ‘he’ intentionally).
So, I am less a believer in being led by a hypothetically high-order being than establishing a work culture based on well-defined and measured principles that is largely self-regulating, and wherever regulation is required — like hiring, firing and pay, for example — a dispassionate AI capable of weighing 183 interdependent variables without our inescapable human cognitive biases should be employed.
3. But people are creative, and machines aren’t.
Leaving aside the creativity of AIs playing Go better than people can, I’ll just say, yes, people are creative. And let’s free them up to spend more time thinking creatively, instead of looking over the weekly TPS reports. But ‘creative management’ is almost an oxymoron.
4. But the idea of driverless management — being led by AIs — is creepy.
People thought the idea of riding in a train or a plane was creepy, too, before it became as common as dirt. Yes, some will make Skynet¹ or Daemon² jokes, and others may worry about biases built into the reasoning of the AIs. And we can lament the fact that a lot of people with the title ‘manager’ will have to go back to being contributors instead of managing others.
But the evidence against human management of humans is overwhelming. We’ve accepted it because we’ve had no other choice, just like humans driving cars. We‘ve become habituated to the stupidity of 30,000+ people killed in car accidents in the US annually, and the enormous costs of cars that sit idle 95% of the time. And we’ve become inured to the huge costs — financial, emotional, and societal — of the mismanagement build into the wiring in the world of business.
Algorithmic HR and rules-based self-management approaches like Holocracy will make driverless management seem like an obvious next step, and not a revolution. But it is.
Driverless management may seem like am empty, shameless effort to piggyback on a the exploding discourse around driverless vehicles. But I think the connection between the two is more than coincidental. They are coeval, derived from the same opportunities and economic motivations. Just as the individual, the neighborhood, and the world will benefit from driverless transport, the worker, the company, and the economy will benefit from driverless management.
And, to close with the original theme: driverless management does not make everyone a manager, any more than driverless cars make everyone a driver.
This transition has already begun, right under our noses. Just as automated lane changing and parking set the stage for fully autonomous cars, algorithmic HR and rules-based self-management approaches like Holocracy will make driverless management seem like an obvious next step, and not a revolution. But it is.
- The malevolent AI in the Terminator series.
- The benevolent — sort of — AI in Daniel Suarez’ Daemon series.