Six Thinking Hats

In the AI Story

De Bono designed the framework for group meetings, where the tangling of modes is most destructive. In a conventional meeting, one participant is being cautious (black hat) while another is being creative (green hat) while a third is reacting emotionally (red hat) while a fourth is trying to establish facts (white hat). The result is not dialogue but collision. The hats synchronize the group: everyone wears the same hat at the same time, so caution happens together, creativity happens together, facts are established together. The modes cooperate instead of colliding.

The framework applies to AI collaboration with startling precision because the AI collaboration suffers from exactly the tangling problem the hats were designed to solve. Most interactions with a language model mix modes in a single cognitive moment: the builder prompts, the AI responds, and the builder simultaneously evaluates accuracy (white), feels whether it sounds right (red), identifies flaws (black), looks for value (yellow), and decides what to do next (blue). The muddled judgment produces muddled collaboration. Separating the modes — directing a dedicated white-hat phase, then a dedicated green-hat phase, then a dedicated black-hat phase — transforms the interaction.

The green hat is the phase most builders skip, because the AI's default output is already competent enough to feel sufficient. The green hat rejects competence. Before evaluating what has been produced, the builder demands alternatives — and this is the phase where provocation, random entry, and other lateral operations live. The AI cannot put on the green hat by itself. The self-organizing dynamics of its training pull it toward the center regardless of instructions to 'be creative.' The green hat is the builder's contribution, and it requires external provocations that disrupt the default pattern.

The red hat has a novel function in the AI age that de Bono could not have anticipated. The red hat gives permission for the intuitive signal — the feeling that something is wrong even when the analysis says it is right. Segal's description in The Orange Pill of the Deleuze passage that 'sounded right' but turned out to be fabricated is precisely a red-hat signal suppressed by white-hat data. The builder who honors the signal and investigates catches fabrication the builder who suppresses it misses.

De Bono's estate recognized the natural fit between the framework and AI when it licensed the creation of a Six Thinking Hats GPT — an AI designed to prompt humans through hat sequences rather than generating answers. The design philosophy inverts the standard interaction: instead of the human prompting the AI for output, the AI prompts the human for thinking. The machine becomes a facilitator of human thought rather than a substitute for it.

Origin

De Bono published Six Thinking Hats in 1985 after developing the framework through corporate training sessions in the late 1970s and early 1980s. The book became one of the most widely deployed management tools in history, translated into dozens of languages and adopted by organizations in over forty countries. A 2024 licensing agreement between the de Bono estate and a major AI developer produced the Six Thinking Hats GPT, bringing the framework into direct engagement with the AI moment.

Key Ideas

Separation as discipline. The framework's power lies not in the hats themselves but in the separation they enforce — one mode at a time, fully, before switching.

Sequence matters. Blue first (define the task), white next (establish facts), green before yellow (generate before evaluating), black near the end, red throughout.

The green hat is the lateral hat. Protected space for creativity, where provocations and random entries operate without immediate critique from the black hat.

The red hat as fabrication detector. In AI collaboration, the intuitive signal that something is 'too smooth' is often the earliest indicator of polished falsehood.

Inversion for the AI age. The machine as facilitator of human thought (prompting the builder through hats) rather than generator of answers — de Bono's late-career vision realized.