Crisis before the revolution

I use AI deeply enough that it has changed the texture of my working day. It helps me move through code, research, language, and decisions at a speed that would have felt implausible only a few years ago. Yet the most important thing I have learned from using it is social. Watch a room when AI enters the conversation and its temperature changes. People become careful about what they reveal, jokes acquire an edge, and a discussion about a tool starts carrying the weight of a discussion about who will still matter.

Leadership makes that tension difficult to dismiss because every organization rests on an unwritten bargain about competence. People learn which forms of knowledge earn authority, which tasks justify a salary, and which years of practice grant someone the right to speak first. Generative AI has begun to interfere with that bargain. It compresses work that once signaled mastery, gives unfamiliar people access to fluent output, and makes private experimentation more consequential than public credentials. The disturbance reaches through job descriptions into identity, status, and belonging.

Psychology gives the reaction a recognizable structure. People are motivated to preserve the legitimacy of systems that tell them where they stand, especially when those systems provide a stable account of effort and reward (Jost & Hunyady, 2005; Jost, Banaji & Nosek, 2004). A capability that competes for resources while disturbing the symbolic meaning of skilled work can produce anxiety, hostility, and defensive judgments (Riek, Mania & Gaertner, 2006). AI anxiety itself spans fears about learning, job replacement, sociotechnical blindness, and the behavior of the systems, so the unease carries material and existential dimensions at once (Wang & Wang, 2022).

Looking back at the history of scientific revolutions offers one lens for this period. A settled paradigm gives a community shared rules for valid problems, legitimate methods, and credible participants. Anomalies first appear as exceptions that the existing order expects to absorb, then accumulate until the repairs become harder to defend and confidence in the old map begins to loosen (Kuhn, 1962/2012). Crisis occupies the unstable interval when the official framework still governs the institution while lived experience has started outrunning it. That interval is emotionally expensive because people can feel a structure losing explanatory power before they can imagine what will replace it.

Modern knowledge work has lived inside a durable paradigm of scarcity. Analysis took time, polished language took time, technical production took time, and those constraints helped institutions sort authority, wages, and trust. AI is puncturing several of those scarcities at once. The resulting anomaly is larger than faster drafting or cheaper code. A machine can now participate in the signals by which people have recognized intelligence and allocated status, which forces every institution to reconsider what judgment, authorship, and expertise mean when fluent production becomes abundant.

I feel the pressure most clearly when capability collides with identity. Being able to survey the literature of an entire field in a single sitting is daunting because it compresses work that institutions taught us to price through time, access, and specialized memory. Accepting what the output demonstrates can be socially expensive. People who controlled access to knowledge must confront a weaker gate, while smaller teams can test claims and enter contests that once required corporate scale. The warnings that follow combine legitimate concerns about reliability, privacy, labor, and concentrated power with an institutional instinct to preserve existing advantage. Those motives often share the same language of safety and responsibility, so careful stewardship and status defense become difficult to distinguish. Evidence then enters a conversation already charged by fear, with each side hearing a different threat in the same result. Work supplies achievement, recognition, belonging, and the esteem through which people approach their sense of potential (Maslow, 1943). When professional output has become a proxy for personal worth, a machine that multiplies output can feel like a subtraction from the self. Nearly everything that once made knowledge work expensive is being compressed at once, and that compression is frightening because the old prices helped people understand their place.

For leaders, this is a test of perception before it becomes a test of policy. Treating the reaction as simple resistance misses the identity structure underneath it, while treating every new capability as destiny abandons the people who must carry the transition. Leadership in a crisis period means naming which forms of judgment remain scarce, rebuilding incentives around them, and giving people an honest account of what the institution now values. Silence leaves status fear to write the story, and status fear usually writes a harsher one.

The people who use AI most intensely may feel the discontinuity earliest because daily contact removes the comfort of abstraction. I can see what these systems still get wrong, and I can also see how quickly the boundary moves. That combination produces a peculiar clarity. Escalating crises often precede a revolution because an old order becomes loudest when its assumptions stop feeling inevitable. We are building toward such a break now. You can feel it in guarded conversations, see it in institutions rewriting their scoreboards, and recognize it in the widening distance between how work is officially described and how it is already being done.