Blog / AI

AI

What the Courtroom Will Look Like

AI is not replacing lawyers — it is changing the ecology of the courtroom, and the species that will struggle are not the ones you expect.

Daniel Jennings · October 3, 2025

The argument about AI replacing lawyers is the wrong argument. It frames the question as a labor market question — who keeps their job? — when the more consequential question is an ecological one: what changes about the environment inside which legal reasoning happens?

Ecosystems do not replace species cleanly. They shift the conditions under which certain traits are advantageous. When a new predator arrives, the species that suffers first is often not the obvious prey but the one that occupied the same niche by a slightly different method. The arrival of a language model in a law firm does not primarily threaten the partner who litigates — it threatens the associate whose job was to pattern-match across thousands of precedents to surface the handful worth reading. That is not a job that disappears; it is a job that gets automated, and the human who did it now has to become something else or become redundant.

What that something else is, is the harder question. The honest answer is that nobody knows yet, because we are at the stage of the transition where the new niche has not been clearly defined. What is visible is the pressure: cases are moving faster; clients expect more preparation at lower cost; the kind of memorisation and procedural fluency that was the mark of a senior practitioner in 1995 is no longer distinguishing. What is emerging, tentatively, is that the trait under selection pressure is something closer to judgment under uncertainty — the capacity to reason about a situation where the precedent is genuinely sparse or contradictory, where the model's output is confident and wrong, and where someone in the room has to notice.

That is a harder thing to train. The law schools have not caught up. Most of the training the profession gives its new members was designed for the ecosystem that existed before the shift, and it does not transfer cleanly. The books GrassMan is publishing in this space are not optimistic or pessimistic about that — they are trying to be accurate, which is the harder stance to hold.

Related titles from this theme

AI.

Catalog titles curated into the same theme bucket on /reading-lists.

Keep reading

More from the press.

Return to the blog for more editorial essays, or enter the catalog.