The AI Friction Principle
AI removes the friction that built leaders’ judgement; the discipline is to put it back, by design, in three places.
​
Learning Requires Friction is the practice of re-engineering productive friction back into AI-assisted work, so that the leader keeps building the schema they would otherwise quietly stop building. This was named and refined through ChangeSchool’s work with senior leaders across our executive education programmes, as part of the modular blocks of our Deeper Work with AI curriculum.
​
The framework
Three patterns, each one a place where AI has removed a step that used to do the schema-building work. The pattern is the same in each case: friction first, AI second. Reverse the order and the friction never happens, because there is nothing left for it to act on.
​
Learning Requires Friction, an inverted-U curve. X-axis: friction in the working session. Y-axis: learning outcome and schema built. Low-end plateau (AI-skim default removes extraneous and germane load alike); middle band shaded as the productive-friction zone; right roll-off (frustration, abandonment risk).
The schema is built in the staring, the recalling, and the reaching of the wall. AI is then applied to an artefact that already carries the leader’s thinking, rather than substituting for it.
​
Two failure modes flow from skipping the friction. Silent decay, surface quality holds while the schema underneath erodes. Illusion of competence, reading a fluent answer feels identical, from the inside, to having thought it; the gap shows up only when a problem arrives that AI cannot help with.
​
Genesis
Robert and Elizabeth Bjork at UCLA named desirable difficulties (Bjork, 1994): conditions that feel inefficient in the moment and turn out to be the mechanism by which durable schema is built. John Sweller’s cognitive load theory (1988) splits task effort into intrinsic (the material’s unavoidable difficulty), extraneous (wasted effort), and germane (the effort that builds schema). AI is good at removing extraneous load, and, used carelessly, equally good at removing germane load. The expertise reversal effect (Kalyuga, Ayres, Chandler and Sweller, 2003) sharpens the picture: worked-example support speeds a beginner and slows the experienced practitioner, because the support does the work the expert needs to do for themselves. AI is the most powerful general support most leaders will encounter; the reversal effect predicts that its uncritical use hurts the leaders whose judgement matters most.
​
The three moves
Pre-draft before prompting. Pick one substantive piece this week — a memo, a difficult email, a diagnosis. Before AI sees it, write the rough version yourself. The pre-draft does not have to be good; it has to be yours.
​
Recall, then check. Pick three moments when you would normally have asked AI a factual question, and write down what you think the answer is before asking. The gap between recall and right answer is the highest-yield information about your own knowledge you will get all week.
​
Run to the sticking point. Pick one problem you would normally have handed to AI from the start. Work it until the thinking actually fails, not until it gets uncomfortable. Then hand it over, with the wall described.
​
How ChangeSchool applies it with executives
We run senior teams through the Friction Audit in three-hour cohort sessions. Bring a recent piece of AI-assisted work and reconstruct two transfers: what the leader learned from AI, and what AI learned from the leader’s standards and corrections. Productive friction is where both transfers happen; skim use registers neither.
​
The discipline
A friction day, one day a week in which AI is paused for a defined slice of work. One category, on one named day, done unaided, to keep the muscle of unaided thinking warm.
​
A recall-test rhythm, three recalls a day in a single line: what you thought, what was true, the gap.
​
A learning budget, a protected block per week, calendared and defended, in which the work is friction-rich by design. A problem worked from scratch; a book read with notes; a conversation prepared for without AI.
Run for a quarter, the three habits become the leader’s defence against silent decay and the illusion of competence.
‘AI removes the friction that was, all along, the mechanism by which leaders built judgement; the discipline is to put it back.’
​
​
​
​
​
Viren Lall, Managing Director,
ChangeSchool LDN (2026).
virenlall.com/ai-friction-principle
​
​
​
​
​
​
AI for Leaders.
Executive Education that changes practice.
Viren Lall is Managing Director of ChangeSchool LDN, a London-based executive education partner. ChangeSchool specialises in AI for senior-leader development, winning the EFMD Global Excellence in Practice Award in 2023 and 2025, with programmes in 39 countries.
​
Since April 2024, ChangeSchool LDN has been designing and delivering mindset shifts through Executive Education Programmes across sectors such as deep tech, manufacturing, and education, for business owners, governance professionals, and senior leaders. Leaders gain AI fluency, protect decision quality, spot value creation opportunities, and foster human-centric AI use. AI capability for senior leaders is also a core element and a constant spine of our Open Programmes for Chief Digital Officers, Chief Operating Officers, and Chief People Officers, delivered by our partner business schools.
Some of our clients include the Royal Academy of Engineering, Education and Training Foundation, and the UK Government's Meet Smart programme.
For speaking, programme, or partnership enquiries, get in touch with him through ChangeSchool LDN.