AI Has Already Killed Academia as We Know It — Inside the Volume Game

AI has already killed academia as we know it, argues a tenured professor and editor-in-chief. The core problem: academia runs on maximalism — most grants, most papers, most students. AI makes volume essentially infinite. The game no longer makes sense.

Student assignments are the obvious casualty

Any take-home assignment is highly likely to be AI-generated or AI-refined. Current detection catches only sloppy users — obvious ChatGPT formatting, comma-separated three-item lists, hallucinated citations, hyperbole. But a student with two paid accounts (e.g., Claude and ChatGPT) who has one AI draft and the other critique/refine, looping until clean, produces work that is not only undetectable — it's better than most human submissions. The system now penalizes the honest student who wrote their own flawed essay and rewards the sophisticated (and higher-spending) AI user.

Research publishing is already flooded

Mass-produced, publishable content is already here. A researcher combining pro subscriptions to Consensus and Claude can generate review articles, methodology pieces, theoretical syntheses, reports, and secondary analyses at a rate close to a paper per day (slowed only by online submission clunkiness). Their CV will quickly eclipse anyone doing independent intellectual work. The same applies to grant submissions — a team of five can run ten applications into a single cycle by rotating the nominated principal investigator (each can submit 2 per cycle). AI is genuinely good at fixing common critical errors: budget flaws, missed citations, eligibility flags. What happens when the number of applications triples?

Detection is already failing

Sophisticated AI use loops between multiple models for critique and refinement, nails formatting and punctuation, double-checks references. This output passes detection and earns higher grades. The rational student maximizes AI use.

Takeaway

The academic game — built on volume of independent human writing — is broken. AI makes it trivially easy to generate undetectable, publishable content at scale. Institutions have not yet grappled with the implications for grading, tenure, and research integrity.