We understand tools through control. We pick them up, give them a task, use the result, and put them down again. The hammer does not disagree. The spreadsheet does not question the assumption behind the forecast. The GPS does not ask whether you are heading to the right place. The user remains intellectually alone.
Most students approach AI in exactly the same way. They ask for an answer, a summary, a structure, or a polished submission. This is hardly surprising. Business schools still reward finished outputs far more than the thinking that produced them, and the race for grades makes outsourcing rational.
AI becomes far more interesting when the relationship shifts from tool use to partnership in thought.
When there are mentally two of you, ideas collide. Arguments are tested. Weak assumptions become visible. One question produces a better one. The value no longer sits mainly in the output, but in the dialogue that changes how the human thinks.
I have tried to create the conditions for that relationship through continuous assessment, progressive incentives, extensive personalised feedback, and repeated coaching. Students must develop their work over time. Simply generating a final answer becomes less useful, while using AI to challenge, improve, and rethink becomes more valuable.
The key is decision. Students who outsource cannot explain why a path was chosen, which alternatives were considered, or what trade offs were made. That is precisely what I reward: judgment, justification, and ownership of the thinking process.
Business schools should redesign assessment around visible intellectual development, not merely polished delivery. Students should submit parts of the dialogue, explain what changed, defend what they rejected, and show how their judgment evolved.
This is where genuine partnership reveals itself. A finished answer may conceal whether any meaningful thinking took place. The conversation behind it shows whether the student challenged an assumption, resisted an easy answer, explored another direction, or changed their mind for a good reason. It exposes the difference between asking AI to complete the work and allowing it to deepen the thinking.
But real partnership requires trust, friction, patience, and depth. It is not created through one clever prompt, nor can it simply be commanded into existence. It develops through repetition, experience, disagreement, and the gradual accumulation of trust. Thus the ground has to be fertile, which is why business schools must create assessment systems that encourage the relationship to deepen, rather than relying mainly on guardrails, constraints, and policing.
If students treat AI as something to exploit or abandon the moment it becomes difficult, they will reproduce the same weaknesses visible elsewhere: shallow thinking, easy cheating, disposable relationships, and giving up too quickly.
The lesson is bigger than AI. We need to teach students how to stay in the conversation.
Interesting. And a big shift from the original "pitch" for AI in education.