No one is as smart as everyone.

That sentence explains more about artificial intelligence (AI) than the thousands of pages colleges and universities have produced trying to define it. AI is not intelligent in the way Newton was intelligent, or Orwell was intelligent. It is intelligent in the way a civilization is intelligent. It remembers what none of us can remember because it remembers all of us at once.

For centuries, the university rested on a simple assumption: knowledge is scarce, and institutions exist to preserve, certify, and distribute it. Libraries were fortresses.

Expertise was difficult to acquire. Professors possessed information that students did not. Degrees signaled access to a world that remained closed to most people.

As higher education expanded, publication increasingly became less a record of discovery than a currency of professional advancement. Papers certify careers, grants, promotions, and institutional prestige. The result is an explosion of research output without a corresponding explosion of transformative ideas.

Novelty has always been rewarded, but only when it is legible within shared methods, shared language, and shared standards of evidence. On this view, AI does not dissolve the institution so much as accelerate it—making fluent, norm-compliant prose cheap while leaving untouched the deeper filters that decide what counts as knowledge.

AI succeeds because academia increasingly rewards the kinds of outputs AI is designed to produce. Which means humans are no longer the smartest secretaries on the planet.

Colleges and universities have responded as institutions often do when confronted with a structural change: by pretending the structure remains intact. We debate whether students should use ChatGPT, whether faculty should disclose AI assistance, whether journals should require new policies. These are administrative questions. The deeper question is why academia was so vulnerable to automation in the first place.

For too long, academia has rewarded those who describe the shadows on the cave wall rather than those willing to leave the cave.

AI was trained on precisely this world.

The machine was built for this game because we wrote the rules.

That is why I am less worried about students using AI to write essays than professors using AI to write papers that are indistinguishable from the papers they already write. The danger is not fraud. It is normalization. Once a system rewards recognizable discourse, a machine designed to generate recognizable discourse becomes an ideal participant. Peer review has always mistaken fluency for thought. AI simply removes the labor required to manufacture fluency.

The coming flood of academic prose will be grammatically impeccable, exhaustively referenced, theoretically aligned, and increasingly difficult to distinguish from entirely human work. The famous Sokal affair fooled one journal. AI raises the possibility of fooling an entire culture of evaluation by producing exactly the kind of writing that culture has spent decades selecting for.

This is not because AI is a genius. It is because academia accidentally optimized itself for predictability.

Creation is not elegant prose. It is the willingness to pursue an idea that may fail. Evaluation is not choosing between two polished paragraphs. It is deciding which questions deserve to exist before anyone knows the answer.

Scholarship is no longer the production of knowledge—it is the policing of its boundaries; Synthesis is not combining fifty articles into a seamless literature review. It is seeing a connection that is invisible only because everyone else has accepted the boundaries of the field; The real crisis, then, is not technological. It is institutional.

Universities still behave as though polished discourse is evidence of scholarship. In an age when polished discourse can be generated on demand, that assumption becomes impossible to defend. Once beautiful prose becomes free, prose itself loses value as a signal of intellect.

The machine was built for this game because we wrote the rules.

What remains scarce is not information, and perhaps not even intelligence. What remains scarce is the courage to think something that cannot be predicted from what everyone already believes.

The future of higher education will not be decided by AI. It will be decided by whether universities can learn to insist on discovery when their systems are designed to reward consensus.