The Pattern Trap

In the AI Story

De Bono was relentless about the point and made himself unpopular with precisely the audiences most in need of hearing it. Expert committees, he argued, produce the most conventional solutions because the patterns of expert training overlap in established territory and diverge in unconventional territory. The group gravitates toward overlap — the territory where all experts agree — and this territory is, by definition, the most conventional available. The expert committee produces an expert-level first idea and refines it with expert-level vertical thinking without ever noticing the framework, because questioning frameworks is not what experts are trained to do.

The AI version of the trap is more dangerous because it is more fluent. A language model trained on software engineering knows the patterns of software engineering with computational fluency. The output follows best practices, handles edge cases, conforms to conventions. It is — in de Bono's precise phrasing — the first idea at expert scale. Not the naive first idea of a novice but the highly refined first idea of collective expertise, which is deeply trapped.

The bilateral trap is what makes AI collaboration specifically vulnerable. Expert plus expert-pattern-AI produces a double convergence: two pattern systems reinforcing each other's defaults. The collaboration feels productive, the output is polished, and the work proceeds with the satisfaction of shared framework. Nothing about the experience signals the trap. The builder who wants to escape must bring a tool the collaboration cannot generate from within itself.

The escape route is metacognitive awareness — thinking about one's own thinking, noticing the framework being operated within, identifying the assumptions the framework makes. Self-organizing systems cannot perform this operation from inside. The brain cannot see its own channels; the model cannot see its own distribution. De Bono's tools (Six Hats, provocation, PMI, CAF, OPV) are collectively a technology for making the invisible pattern visible so that the lateral move becomes possible.

Origin

The pattern trap concept runs through de Bono's entire corpus, with distinct articulations in The Mechanism of Mind (1969), I Am Right — You Are Wrong (1990), and the later creativity works. It has analogs in Thomas Kuhn's paradigm theory, Chris Argyris's governing-variable framework, and Byung-Chul Han's critique of convergent aesthetics — each emphasizing different aspects of the same structural phenomenon.

Key Ideas

Expertise as imprisonment. The deeper the channels, the harder the lateral step — a structural property, not a personal failing.

First idea at expert scale. AI reproduces the default-pattern problem that de Bono identified in human experts, but with superhuman refinement of the default.

Bilateral convergence. Human expertise plus AI pattern-following produces the most polished, most conventional output available — the trap at double strength.

Metacognitive awareness as escape. Making the framework visible is the precondition for stepping outside it; self-organizing systems cannot do this from within.

Tools are technology for visibility. PMI, CAF, OPV, Six Hats, provocation — each makes a dimension of the invisible pattern visible, enabling deliberate disruption.