AI Is Getting Its Eternal September as Every Field Repeats the Same Introduction

Ethan Mollick has named a side effect of AI's rapid spread: every smart commentator approaching it from another field has to pass through a similar set of foundational questions. The result may be an Eternal September for debates about AI safety, consciousness and jobs.

Politics, culture and economics are entering the same debate at different times

Mollick's post follows an earlier observation that today is the least we will ever hear about AI. He now adds that commentators in politics, culture and other fields will catch up on the same AI topics while writing about them.

This is not a product announcement or a new research result. It is an observation about how knowledge spreads. AI is entering more specialist conversations, but participants do not arrive at the same moment or with the same vocabulary.

An expert in one field can still be a newcomer to AI

That changes how public commentary should be read. A political scientist may understand institutions deeply while reopening a familiar question about anthropomorphizing models. An economist may offer a precise labor-market analysis while still learning the difference between a benchmark and system behavior in real deployment.

Repetition is therefore not merely noise. New domain experts bring AI into contexts the technical community often understands less well. Value appears when they connect their field knowledge to existing evidence about models instead of running the two conversations separately.

A metaphor from 1993 warns of lost collective memory

Eternal September originally described the influx of new users to Usenet 33 years ago, around 1993. Communities had been accustomed to teaching newcomers their norms each September. Mass access turned that seasonal cycle into a permanent condition.

AI can produce a similar mechanism: the same dispute returns with a new audience and little memory of previous arguments. Experienced participants may mistake fatigue for evidence that a question has been settled. Newcomers can mistake an old argument for a discovery of their own.

Bridges between domain expertise and AI literacy will determine quality

The useful signal will be newsrooms, universities and professional bodies that establish a short common foundation, then let experts focus on what they know best. Links, definitions and the history of disputes will matter more than another round of general panels about the future of AI.

Mollick's observation is ultimately about debate infrastructure. Every field needs its own translation of AI's consequences, but it should not have to reinvent the entire introduction. Without shared memory, public conversation will grow faster than its capacity to learn.

Lilith's verdict

Parliaments, newsrooms and classrooms are playing the same first lecture on AI, just on different days. Without shared notes, the rooms will keep filling while the opening slide never changes.