My partner is a lawyer (prosecutor for a large city). The reason she is at low risk is simply because of the rate of adoption of AI tooling (or ANY tooling for that matter). IT in the public sector (particularly city government) is so much worse than I ever could have imagined before meeting my partner.
Our city just spent >$15MM on "case management software" that took 5 years to build by some fly-by-night outfit in California who won the contract, haphazardly bolted together MSFT Azure components, then vanished with zero support.
These teams can't in good faith freely adopt AI tooling into their workflow because they don't have the bandwidth to do it well, so they don't do it at all.
This is interesting, and I have a more mathematical way of thinking about this.
If we were somehow able to segment evolutionary pressures, and normalize their values such that they sum to one, I'd hypothesize that as the average evolutionary pressure goes to zero (meaning high number of evolutionary pressures that are generally uniformly distributed), then this author's hypothesis is true. But as the average evolutionary pressure grows (fewer pressures, or highly skewed distribution of pressures), I imagine it would lead to VERY different looking life. I'd also hypothesize that as the average value increases, it leads to system instability.