What leaders are really saying: They'd rather prompt than lead

I saw this shared on Substack and laughed out loud. Not just because this is really how Claude sounds when it’s failed to produce the requested task, but because it reminded me of a real contradiction that is being made visible in leadership right now.

For the past 18 months, we’ve watched leaders eagerly invest time and patience coaching AI systems: iterating, re-prompting, troubleshooting, etc. And if you pay attention to that, it does highlight what the real driver of motivation has always been for these leaders: compliance. Not collaboration. Not creativity. Compliance.

Leaders decide what is worth their attention based on what is going to be the fastest route with the least friction. AI is a product. It cannot think. It cannot reason. It cannot feel. And it cannot plan. It's designed to be sycophantic. The relentless focus that the modern workplace has put on productivity and efficiency over the past several decades has directly helped to lower individuals’ thresholds for relating to other people's experiences, because it's all about how quickly you can move through your tasks. It rewards the transactional. So when you have leaders who view leadership as a series of transactions, AI provides them with a very attractive container; they don't have to deal with someone else’s thoughts, feelings or opinions. They can just deal with the machine. And their relief is tangible. 

To me, that highlights something I talk about with clients all the time: how corporate America encourages us to disconnect our brain from our central nervous system. If CEOs could tell us to leave our central nervous system at home every morning and never bring it into the workplace, they would.

This data-rich, highly sophisticated organ system is designed to feed us information non-stop. Not just about tactical stimuli, but about the environment in which we are operating. While your brain’s frontal cortex is processing  data on logic and tasks, your central nervous system is also constantly pulsing data that’s driving your motives and behaviors. Society rewards ignoring the latter and shoving all of that data to the side, which leaves us with all of this deeper insight left on the table that reveal the internal narratives informing our decisions.

Some people will frame it as: "People are unpredictable. AI is at least consistent." Which is amusing to me, because AI is not consistent either. That's why it requires multiple prompts, reiterating and redirection.

But what these narratives do is justify for leaders motivated primarily by compliance why it's better to give all their attention to a machine instead of the people who've been selected to do the work. It's an exercise in chasing perfect compliance with minimal effort or interaction, which AI also doesn't really provide. This typically indicates a deep-rooted need for control to feel safe, an underdeveloped ability to tolerate ambiguity, or both. 

This is where I always come back to David Rock’s SCARF model. It maps out the primitive brain's five primary social threat domains. Our amygdala, the primitive part of our brain designed to scan our environment for threats, real or perceived, is only focused on one thing: our survival. It doesn't care if we're happy or content or fulfilled; it just cares  about keeping us alive.

SCARF breaks down those five threat domains: Status, Certainty, Autonomy, Relatedness, and Fairness. The two that matter most in this conversation are Certainty — the need to know what's going to happen next — and Autonomy — the need to control the outcome. Our brains don’t like gaps of information, so it will reflexively insert contextual data, real or invented, that “makes it make sense” to us. For someone with a low threshold for ambiguity, feeling uncertain is perceived by the amygdala as a life or death threat. And in the context of AI, the lure to finally claim total certainty and autonomy becomes a panacea for low frustration tolerance and discomfort. It offers safety and relief in the idea that AI will produce guaranteed perfect output every single time. Although so far, there's little evidence that's what's actually happening.

This connects directly to how leaders hold power, because there are three kinds of it: positional, expertise, and relational.

The most important kind of power is relational, because that's where influence lives. Building connection, building reputation, building community that fosters collaboration because everyone has skin in the game. When leaders trade-off this off in favor of a command-and- control leadership style, they struggle to build trust, respect, and credibility with teams. AI gives them permission to relinquish their own power to connect, collaborate and inspire others in service of speed and productivity.

I've noticed lately that increasingly, my coaching clients working with AI are asking: "how can I build more charisma?" This is very telling: we’ve gone so far down the road of conflating the performance of charisma with actual competence. You can wear the mask of an entertaining leader and not possess expertise. 

And honestly? I think AI is needy. I think AI is exhausting. One has to be so very granular and specific with AI to coax out the desired result - often requiring more effort when working with real people. But the perception persists that humans are too complicated, too messy, too draining for all that. What leaders really don't want to deal with is the human condition: the friction of unpacking a dissenting point of view, a nuanced question, a complex emotion, because it might elicit something in them too

So many leaders operate from the belief that in order to lead effectively, you must always have the appearance of being right. You cannot ever say "I don't know" or "I was wrong about that” - if anything, That’s the threat of uncertainty and low frustration tolerance rearing its discordant head yet again. Of course, the intent behind wanting to show up prepared is completely understandable. But here the thing - we all understand there’s not one person who knows everything about anything. The most gifted astrophysists see themselves as scholars of the universe, not experts, because the universe is constantly surprising them with new information. They embrace curiosity and intellectual humility as features of their competence, not bugs to be eliminated. When leadership becomes a performance of infallibility, what’s overlooked is development of true expertise - for your people and your own. You're limiting both of those routes for growth.

AI is but a series of mathematical equations all made from our collective IP - our “human spark”. A frictionless path actually leads us to the places we've been trying to avoid: less innovation, less creativity, less tolerance for diversity of thought. All of which close us off to new paths of knowledge.

If I could put one question in front of every organizational leader right now, it would be this:
If you put your calendar and your stated priorities side by side in front of your team tomorrow and did not explain it, what story would they tell about who or what is actually important?

The modern workforce has prioritized efficiency and productivity for decades. Organizations have decided putting all of their chips on AI will finally be the golden key to perpetually exponential profits. What it’s costing you is your own capacity for relational power and influence. You're not actually gaining more productivity — you could be maximizing the productivity of both yourself and your team if you were consistent about investing time, patience, and authentic support for performance development. But when we look at how leaders’ time and efforts are actually being spent and on what, the message is clear. And so is the cost.