AI is triggering a quiet crisis of confidence among knowledge workers, who are withdrawing their own judgment before being asked to leave, threatening the success of AI transformations.

What happened: Workers across knowledge professions are experiencing a 'self-replacement dynamic' where they pre-emptively shrink their sense of agency and contribution in response to AI tools, without being asked to do so by their organizations. Unlike visible resistance, these employees appear compliant and use the tools, but their judgment and willingness to contribute have quietly left the organization.

Why it matters: Three specific triggers accelerate this withdrawal—loss of control (AI is imposed rather than chosen), loss of trust (AI announcements paired with headcount reductions create the impression workers are costs to optimize), and loss of legibility (workers cannot locate themselves or their value in the new landscape). Human judgment is precisely what separates successful AI transformations from failed ones, so this quiet disengagement is more damaging than overt resistance.

What to watch: The pattern mirrors what happened to scribes after Gutenberg—some gradually stopped believing their skills mattered and disappeared from the profession, while others discovered their deep knowledge was indispensable in new roles like editors and curators. The shift from self-replacement to AI-forward often starts from a single moment where a worker discovers something only they bring to the table (their judgment, taste, or understanding of edge cases), but these moments 'rarely a training event that can be scheduled or mandated.'