AI-Compared psychology analysis, cladistic parcimony and Clever Hans. Commentary to the article by Jonathan Birch and Kristin Andrews.

The article by Kristin Andrews and Jonathan Birch examines the question of AI sentience through the lens of comparative psychology and cladistics. They introduce the “gaming problem,” suggesting that AIs can mimic human behavior without actually being sentient, which could mislead observers into attributing sentience to them. In assessing sentience, they seem to prioritize evolutionary markers over experimental ones. I offer two main critiques: first, their “cladocentrism”, which excludes non-biological sentience from the outset; and second, I draw a parallel between their “gaming problem” and what I call a “Hans problem”, inspired by the case of Clever Hans, showing that this issue of trickery also arises in comparative psychology.