Why AI Training Won’t Fix Your Leadership Problem
- Aug 14
- 5 min read
Updated: Aug 19
Jivi Saran is globally recognised, for advancing Quantum Business and Conscious Capitalism. A Senior Business Advisor, Scholar, and Best-Selling Author, Jivi blends rigorous research with 35 years of executive advisory experience to elevate leadership and business transformation.
Boards are approving AI training budgets faster than they can spend them. Executives are learning the tools, the models, the prompts. Almost none of it touches the thing that decides whether an algorithm sharpens a decision or quietly takes it over. Research with senior financial leaders points somewhere less comfortable than a training gap.
What happens in the moment a leader defers to the machine?
It’s four o’clock on a Thursday, and the recommendation is on the screen. The model has chewed through more data than anyone in the room could hold. It’s confident. It’s fast. It’s also, in this case, optimizing for something nobody has said out loud.
The executive feels a flicker. Something about it doesn’t sit right, not a numbers problem, but a values problem. Then the meeting moves on.
That flicker is the whole thing. Whether a leader acts on it or lets it slide has almost nothing to do with how much they know about AI. It has everything to do with how far they’ve grown as a leader: whether they can catch their own reaction, question it, and treat their own judgment as being worth as much as the machine’s.
There’s a name for what happens when that ability is missing. Brock and von Wangenheim (2019) call it automation bias, our habit of trusting the algorithm even when our own judgment would have got us to a better answer. Companies treat this as a training problem. It isn’t. It’s a growth problem, and knowing more about AI does nothing to fix it.
Self-awareness isn’t where it starts, it’s what you get
I interviewed 25 senior financial leaders. The theme that came up most often wasn’t values, ethics, or good intentions. It was conscious engagement, actively paying attention, which showed up 177 times. Reflection, the thing that produced that attention, showed up 92 times.
The order is the finding. The leaders making the best decisions didn’t start out self-aware and then act on it. They did the work first, asking themselves hard questions in real time about why they were about to do what they were about to do. Self-awareness came out the other end.
One of them put it perfectly:
“Like a car dashboard is saying you’re misaligned with your values.” – Financial leader, Saran (2026)
The dashboard isn’t the engine. It’s the readout.
That should change how companies spend their development budget. Personality tests, 360s, and strengths profiles all try to hand people self-awareness directly. They go after the result instead of the thing that produces it. What improves decisions is a habit: stopping long enough to notice your own values, assumptions, and blind spots before they make the call for you.
With AI in the room, that gets urgent. An algorithm won’t wait for you to catch up with yourself.
Three kinds of leaders, three ways of handling the machine
Robert Kegan (1994) spent his career mapping how adults grow up, and his map lines up with AI risk almost too neatly. Three positions, three very different outcomes.
The leader who borrows their identity from everyone else. Kegan calls this the socialized mind. They take their cues from what’s expected of them. Put an AI recommendation in front of them and they defer to it, then look around for agreement. Automation bias runs free here because there’s no independent view to check it against. This is the riskiest place to be, and there are more senior people sitting in it than most companies would like to admit.
The leader who has their own compass. The self-authoring mind. They’ve built their own set of values and can hold a recommendation up against it. Much safer. The risk that’s left is a leader so sure of their own framework that they stop testing it.
The leader who can hold more than one view at once. The self-transforming mind. They let the machine’s answer and their own judgment argue with each other instead of picking a winner straight away. This is where the best thinking happens. It’s also rare.
Maslow (1970) called the journey behind all this self-actualization, becoming fully yourself and using what you’ve got in service of something bigger than your ego or your employer. It sounds abstract right up until you watch it fail in a room where an algorithm has just recommended something convenient and wrong.
The four things that get in the way and how AI makes them worse
Across a survey of 202 financial leaders, and as confirmed in the interviews, four barriers came up again. They can shut down good decision-making no matter how decent a leader’s values are: time pressure, industry norms, profit demands, and systemic constraints.
Unconscious habits and those four barriers turned up together 411 times in the interview data, the strongest link in the whole dataset. They aren’t two separate problems. They’re one problem feeding itself.
AI makes all four heavier. It shortens the time you have to decide, which kills reflection. It rewards thinking that puts numbers first, which hardens the norms. It optimizes for goals that may quietly work against the people you serve, which sharpens the profit pressure. It adds a new constraint of its own: a recommendation that already carries authority before a single human has looked at it.
The pressure is rising. The question is whether leaders are rising with it.
You can’t build this in the moment
One participant said something that has stayed with me:
“I always had the philosophy that I don’t know it all.” – Financial leader, Saran (2026)
That isn’t humility for its own sake. It’s structural. A leader who hasn’t built the ability to make an independent call guided by values won’t magically find it the day an algorithm suggests something convenient. It must already be there when the test comes.
Four things build it. None of them need a new platform.
Put reflection in the calendar. Reflection that depends on spare time never happens because there is no spare time. Book it like a board meeting.
Ask what the system was told to optimize for. Not just what it recommended, but what it was aiming at and who benefits. Most executives have never asked.
Treat hesitation as information. That flicker of discomfort is often the only warning that something is off. People need permission to say it out loud and practice doing it.
Practice judgment when nothing is riding on it. Judgment built under pressure is built too late.
The skill nobody is budgeting for
Knowing how to use AI is table stakes, and it will be cheap and common soon enough. The ability to think for yourself in front of it won’t be. You can’t buy it, license it, or install it in a workshop lasting two days, which is exactly why it will separate the companies that use AI well from the ones it slowly hollows out.
If you’re building an AI plan for your executive team this year, ask it one question: does this develop the people who will have to disagree with the machine?
If this raises questions about your own leadership team’s readiness for AI-assisted decision-making, I would welcome the conversation.
Read more from Jivi Saran
Jivi Saran, Conscious Capitalism Scholar & Practitioner
Jivi Saran is a transformative business advisor, scholar, and thought leader whose work bridges quantum principles, human consciousness, and organizational strategy. With over 35 years of guiding executive teams, she empowers leaders to make purposeful, future-shaping decisions that elevate both performance and humanity. As the founder of Quantum Business Growth and author of Quantum Business: Leading with Soul in a World of Systems, Jivi champions a new era of leadership grounded in clarity, coherence, and conscious capitalism.
