The Intelligence Trap

In the AI Story

The trap operates through a specific mechanism: deep patterns of expertise make vertical reasoning faster and more accurate within the framework, which reinforces confidence in the framework, which makes the framework harder to question, which deepens the patterns further. The expert's pattern is the expert's cage. The channels that make expertise possible are the channels that make lateral escape from expertise structurally impossible — not difficult, not unlikely, but structurally excluded by the optimization the expertise represents.

AI has made the intelligence trap computationally visible. A large language model trained on expert corpora inherits the patterns of expert framing. Its output is expert-level; it is also pattern-bound. The model can produce software architecture that follows best practices, handles edge cases, and conforms to conventions with a fluency no individual expert could match. This is — in de Bono's precise terminology — 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 especially dangerous without lateral tools. The AI follows the pattern of its training. The expert builder follows the pattern of their expertise. When the two collaborate without intervention, a double convergence occurs — two pattern systems reinforcing each other's defaults, producing output more polished and more conventional than either could produce alone. De Bono observed this in expert committees and called it 'group think.' Scale it to human-plus-superhuman-pattern-follower, and the convergence becomes nearly irresistible.

The escape from the trap is not more intelligence. It is a different cognitive operation — the lateral move that steps outside the framework rather than drilling deeper within it. The tools for the move are not gifts of genius; they are specific techniques (provocation, random entry, the Six Hats) that disrupt the pattern through mechanical interventions any practitioner can apply. The skill is not beyond the expert's reach; it is simply not what expert training has rewarded.

Origin

De Bono articulated the intelligence trap across several works, most directly in I Am Right — You Are Wrong (1990). The concept draws on his self-organizing pattern theory — intelligence is the depth of the channels; the trap is what the depth excludes. The term has since been borrowed by David Robson (The Intelligence Trap, 2019) and others, though often with frameworks that diverge from de Bono's mechanical account.

Key Ideas

Intelligence ≠ creativity. The capacity to reason powerfully within a framework is structurally different from the capacity to change the framework.

Expertise as cage. Deep patterns make expertise possible and lateral escape impossible — the same channels perform both functions.

Bilateral convergence in AI collaboration. Expert plus expert-pattern-AI produces the most polished, most conventional output — the trap at double strength.

Escape is mechanical. The tools for stepping outside the pattern are specific techniques, teachable and practicable, not gifts of insight.

Diagnostic mark. If the AI collaboration feels like refinement, the trap is probably active; if it feels like disruption, a lateral move is in progress.