Tuesday, 8:40 a.m., a third-floor meeting room at an insurance company in Lille. On the screen, a 28-year-old data scientist runs her demo: an AI tool sorts incoming claims, spots the simple files and routes them to automatic processing. Pilot result: two days cut from the handling time of each file.
Simon raises his hand. He has run the claims department for nineteen years.
“Our process is heavy, I’ll grant you that. But it’s reliable. It has proven itself.”
Around the table, people nod. Nobody mentions one detail: the week before, at the coffee machine, Simon was cursing that very process. “Three sign-offs for a windshield. It’s insane.”
Simon is neither stupid nor a hypocrite. He has just done exactly what social psychology has described for more than thirty years under a slightly clunky name: system justification.
Defending the system that weighs on you
In 1994, psychologists John Jost and Mahzarin Banaji published a paper with an unsettling hypothesis. We need to see ourselves in a favorable light, and to see our group in a favorable light. Jost and Banaji added a third need: to see the system we live in as fair and legitimate. Under some circumstances, that need would override our personal interest and our group’s interest (British Journal of Social Psychology).
The theory first targeted stereotypes, and the way disadvantaged groups end up adopting a negative image of themselves. It later expanded to judgments of fairness, merit and legitimacy, all the way to political ideologies. Jost gives it a palliative function: defending the established order makes the situation easier to bear. He sees it as an answer to deep needs to reduce uncertainty, threat and social discord (Sapienza).
Reread those needs with Simon in mind. The new tool creates uncertainty: how will we work tomorrow? It looks like a threat: what happens to my department? And criticizing it in front of everyone would put him at odds with his team, who built their skills on the old process.
The Decision Lab offers an example I like because it has nothing to do with business. A student tells a teacher that he works nights, has no internet at home and can’t afford tutoring. The teacher replies that he grew up the same way, made it, and the student will too. The encouragement is sincere. Yet it turns structural barriers into a matter of individual willpower, and the system stays intact (The Decision Lab). In your executive committees, the corporate version comes up every week: “I learned on the job, they’ll learn too.”
Before you blame your teams, two serious counterarguments
I owe you a warning, because this theory is often told wrong. The spectacular version claims that the most disadvantaged defend the system even more than the privileged. That’s the version you find in popular articles.
It holds up poorly. In 2013, Mark Brandt of Tilburg University tested it on three large representative datasets, two American and one global, totaling 151,794 people. In most cases, the effect ran in the opposite direction or was close to zero. People with the lowest status do not legitimize the system more than others (Tilburg University).
What remains solid is more modest, and more useful to a leader: all of us are motivated, to varying degrees and depending on context, to defend what already exists.
The second counterargument is more uncomfortable for innovators: the person resisting is sometimes right. Simon’s heavy process may stop three frauds a year that the tool will miss. Before you diagnose a bias, check that you aren’t facing expertise. Your own hypothesis (“they resist because they’re afraid”) is a hypothesis like any other. Kill it with experimentation.
Status quo bias, its lab cousin
System justification has a better-known cousin in behavioral economics: status quo bias. In 1988, William Samuelson and Richard Zeckhauser showed that an option presented as the current situation becomes clearly more popular, with identical content. That advantage grows with the number of alternatives on offer. They found the effect in real decisions too, such as the health insurance and retirement plan choices of Harvard staff (Journal of Risk and Uncertainty).
Does it still hold? In 2021, a team reran four of the original scenarios with just over 600 participants. Three out of four replicated clearly: allocating a budget, choosing an investment portfolio and choosing a job. The fourth, choosing a car color, showed nothing clear (Meta-Psychology). Keep that nuance in mind: the bias weighs heavily when the stakes matter, much less when the choice is trivial.
My brain is no exception, and neither is yours. Faced with a complex decision, it saves energy and picks what already exists. On stage, I run an experiment on this bias: between 85% and 90% of the room makes a decision that matches the status quo, and astonishment spreads across faces when I reveal the mechanism (My book, chapter 6).
The four situations that make your teams rigid
The work of Aaron Kay, at Duke University, and Justin Friesen brings the most valuable nuance for a leader: system justification depends on context. Their 2011 review identifies four situations that trigger it: threat to the system, dependence on the system, the feeling that you can’t escape it, and low personal control (ScienceDaily).
Two experiments make this concrete. Students made to feel dependent on their university approved a funding policy when it came from the university, and disapproved of the same policy when it came from the government. Participants convinced they could not leave their country explained a 20% pay gap between men and women by “natural” differences, instead of blaming the system (ScienceDaily). Kay describes an active process: people mobilize to defend the established order.
Let’s transpose this to business. Fair warning, this is my extrapolation: these studies cover social and political systems, not claims departments.
Threat. Every time you present AI as a matter of survival, you trigger the mechanism that makes people close ranks around what exists. Panic is directly linked to the survival instinct, and the survival instinct protects what it knows.
Dependence. Simon built nineteen years of career on this process. His legitimacy depends on it. Asking him to criticize it in a meeting means asking him to saw off the branch he’s sitting on.
No way out. At fifty-four, in a region where his job is getting scarce, Simon doesn’t see himself leaving. The more stuck you feel in a system, the more virtues you find in it.
Low control. The decision to bring in the tool was made two floors up, without him. The less control you have, the more you cling to the order that reassures you.
All four conditions are present in that meeting room. Simon’s reaction was predictable before the demo even started.
Defusing the reflex: what you can do on Monday
If context triggers the defense of the status quo, context can also defuse it. Here is what I take from it, with the same caution about extrapolating.
Lower the threat before you talk about tools. Say what you’ll do with AI, for what purpose, and what you won’t do with it. A communication vacuum always fills up with the worst-case scenario. Without psychological safety, nobody dares to say the current system no longer works, so everyone defends it in public and criticizes it in private (My book, chapter 9).
Shift the dependence. Simon depends on the process. Make him depend on something else: his expertise on difficult files, which the tool can’t handle. Give him the job of defining the sorting rules. He keeps his legitimacy, and it now attaches to the new system.
Give back control. A pilot Simon designs with the data scientist has a different status from a tool imposed on him. Measure before, during and after, with him.
Move the change from proposed to established. Justin Friesen, Kristin Laurin and Aaron Kay describe our tendency to treat the current situation as the way things should be. They also show that reactions to a change evolve depending on whether it is proposed, imminent or already in place (Duke University). A change that stays forever under discussion feeds resistance. A real, bounded, measured pilot quickly becomes part of the new status quo, and your teams will end up defending it with the same energy.
Finally, rate yourself. On a scale from 0 (subject to it) to 5 (not subject to it), how much do you defend in meetings a process you criticize at the coffee machine? I’m not always at 5 myself. Nobody is (My book).
The essentials
Remember three things:
- Your teams defend what exists to reduce uncertainty, threat and the risk of isolation, even when they criticize it in private.
- The idea that the most disadvantaged defend the system the most doesn’t survive the data; the general motivation to defend the status quo does.
- Threat, dependence, no way out and low control trigger this reflex: work on these levers before you deploy a single tool.
Before you accuse your teams of inertia, look at the context you created. Innovate or agonize, the choice is up to you!
A corporate event, a seminar, an executive committee meeting, or a management committee meeting?
Philippe’s keynotes on innovation and AI are groundbreaking, you’ve been warned!
Complement a powerful keynote with innovative and impactful workshops.
A keynote inspires and raises awareness; workshops transform!
References
- (My book) https://philippeboulanger.com/book/
- (British Journal of Social Psychology) https://banaji.sites.fas.harvard.edu/research/publications/articles/1994_Jost_BJSP.pdf
- (Sapienza) https://dippsi.web.uniroma1.it/it/node/7069
- (The Decision Lab) https://thedecisionlab.com/reference-guide/sociology/system-justification-theory
- (Tilburg University) https://repository.tilburguniversity.edu/items/b4c17f22-0790-472f-9980-7e23c9dc98b8
- (Journal of Risk and Uncertainty) https://scholar.harvard.edu/files/rzeckhauser/files/status_quo_bias_in_decision_making.pdf
- (Meta-Psychology) https://open.lnu.se/index.php/metapsychology/article/view/2470
- (ScienceDaily) https://www.sciencedaily.com/releases/2011/12/111212153157.htm
- (Duke University) https://scholars.duke.edu/publication/1350687







