The Manager In The Machine.
Exploring the fragile line between human leadership and AI orchestration, and why rewriting the social contract of work may be the only way to keep meaning in our workplaces.
The human edge in a technology-driven workplace exists. We want to believe that it will continue to exist in the future world of work.
What if the human edge disappears? What if we can’t justify it anymore? What if the worker twin becomes a boss twin?
Dystopian?
Well, we can debate the ROI of AI. With a 0% return, the MIT was blunt about it recently in this report.
We can challenge its output too, but AI is marching forward. In some organisations, AI is morphing into a twin. For instance, Deutsche Telekom is piloting an AI “digital twin” of its call center workforce to optimise responses and decisions.
And AI-run businesses aren’t merely futuristic. Experiments are testing the boundaries. In one study by Carnegie Mellon, AI agents were tasked with running a fake company. The result? Even the top performer delivered just 24% of assigned tasks, and triaged basic business operations failed across the board. We are not there yet, but it’s moving.
Its next logical evolution would be to become a manager – a not-so-hypothetical leap in my view. Today’s models can already complete tasks. Add a layer of reasoning to connect the tasks to the bigger picture and an AI enterprise system could build up an understanding of the job at hand. A cold, colourless, synthetic understanding deprived of emotions, for sure, but an understanding, nonetheless. Go one step further and AI is now orchestrating a complex ensemble of work-related tasks, paving the way for people orchestration – assuming the latter are still needed.
Beyond profound societal implications, the manager in the machine may not be a desirable outcome for the future of work.
AI trusts you to trust it
Answer this. Who do you blame if AI makes a bad call? Who do you thank if it makes a good one?
As human beings, we have a tolerance threshold built in. In the world of work, we tend to be quick-to-blame and slow-to-praise creatures – DM me if your team shows an opposite habit – but we tolerate flaws in human leaders.
Delivering on a vision, placing calculated bets, communicating strategically day in day out: these are extremely demanding tasks that come with bumps in the road at times. These occasional hiccups make the journey real. As counterintuitive as it sounds, these deviations inspire if they are genuine. They magnify the value of success.
If AI leads, would we be as tolerant? Would we even care about “work” if it is reduced to a series of robotic interactions (pun obviously intended)?
If AI leads, the whole meaning – and value – of leadership comes crashing down. The social contract disappears.
Let us rewrite some of its terms properly before it is too late. Let me give it a modest start.
Rewriting the social contract of work
The workplace is not a machine, but if we continue to itemise work everywhere, it may well look like one. Admittedly, human presence is either unnecessary or suboptimal within certain business functions. There, the right dosage of AI is a no brainer. And it is already happening – with a human in the loop, of course.
Let’s go beyond and challenge functions where human presence remains a necessity today. Even there, technologies like generative AI will play an increasingly significant role. This is precisely where I believe a new social contract of work is required.
In broad strokes, it aims to:
Retain a strong social dimension. It may be an unpopular opinion, but healthily managed companies are still great structures for human capital and potential to evolve productively and collaboratively.
Preserve human management and leadership. Building a relationship with an AI creature may be entertaining to some, it must remain what it is: a pastime. Work should be fun – but it’s not yet a pastime.
In finer strokes, this results in three guidelines:
More meeting:
Check-ins, stand-ups, one-on-ones – even done remotely – help not only to maintain a social link between teams, but to generate new thinking.
Naturally, this is contingent upon all participants willing to be constructive and productive – a strong assumption.
Guess what, these lively interactions may turn out creative and then help teams write…better AI prompts.
More fighting:
The workplace must remain a place of tension and celebration – both significantly contributing to elevate everybody in the process.
Fighting does not mean bullying. It does not mean mocking. But it requires some serious listening skills from all parties involved.
Systematically turning to your favourite AI chatbot for work support may remove human friction, but it is an escape route.
More micromanaging:
Loneliness at work may not creep in when teamwork happens. It quickly does when the team is not around, though.
To counter this, more individual attention becomes a necessity. True leaders will check in on their team members with sincerity; awful ones will do it just to tick a box (and that will not help anyone).
AI might be an elephant in the room, but it is invisible when it comes to restoring or sustaining a feeling of togetherness.
Slow and furious. Fast and steady
Reading this, you may assert these guidelines are antithetical to the holy notion of everything now. That they are in contradiction with constant shareholder pressure to streamline and optimise.
I think they achieve the opposite: they reestablish fluidity, clarity, and transparency. And maybe modestly infuse some honesty in human relationships in the workplace.
Not only that: they even set individual performance for collective success. By removing the focus on work items, we allow for meaningful human intervention. It does not mean not using generative AI. It does not mean not using custom GPTs. On the contrary, it means using them when we give them the green light. In other words, going slow to go fast is the name of the game here.