Probably an obvious point: epistemically that’s an error, but politically it’s probably an indispensable tactic. Say you do an honest and perfectly reliable utilitarian analysis, and find that chimpanzees really should not be used in research; the real substantial medical advances are not worth their suffering. But frustratingly the powers that be don’t care about chimps as much as they should. Your only hope is to convince them that chimp-using research is nearly useless to humans, so that even their undersized compassion for chimps will convince them to shut the research down.
I have a kind of romantic suspicion that nearly all politically active people are like this, that if you could somehow get them alone and sit them down and ask them what they really think, they’d go, “Yes, congratulations Einstein, you figured it out. Of course if we succeed then it’s likely the lives of [some group] will get a lot worse, but, well, omelets and eggs.” And then they swear you in and give you a membership card, because if you’ve gotten this far, then you can also see that they’re justified.
(Nods) that’s really what I was trying to say, yeah.
Also it’s worth an NB that the AAT only applies to epistemic agreement, right? It doesn’t prevent groups from competing over resources: we agree that the pie is tasty, which is precisely why we’re fighting over it. Of course if you’re committed to fighting, then screwing with your enemy’s, and partially-committed ally’s, models of the world is a valid combat tactic.