Cognitive dissonance is a theory developed by Leon Festinger in the late 1950s or early 1960s [Edit: First publication appears to be 1957] and supported1 by a pretty large mountain of research. It posits that we have a mental/emotional aversion to holding two directly conflicting cognitions (thoughts, more or less) at the same time. When this happens, it is intensely uncomfortable and thus is usually resolved by our brains very quickly, without us realizing it.
Cognitive Dissonance has many potential implications but the most commonly studied one (AFAIK) is: we realize we have done something, for whatever reason, that violates our ideas of what we should do, so we change our attitudes about OK/not OK behavior. In Festinger's original experiment participants realized they had just told others a task they had done2 was interesting when it was actually intensely boring; the participants generally changed their minds to actually believe it was interesting.
When our brains experience the two conflicting cognitions:
I do not think I should do XI just did X...the easiest way for our brains to resolve the dissonance is to change cognition 1, because changing cognition 2 implies distrusting our clear memory of "hard facts". The result is that we change our opinion or attitude about the thing we did, usually with no awareness that we changed it or why. If confronted, we'll probably make up a reason that doesn't quite explain our change of heart.
People experiencing cognitive dissonance aren't unsure or "conflicted"; they are certain. Cognitive dissonance explains the otherwise slightly bizarre ingroup devotion in fraternities, sororities, military units, and cults, for example.
Making it all about US Politics
At this point I usually remind people of something Trump does: He gets his toadies to do increasingly awful things for him, tell increasingly transparent lies, violate their own values. I don't think this is accidental; I think it's like a TV mob boss (which Trump is) getting the new guy to murder someone. Now the mob boss owns the new guy, and not principally because he has blackmail; because the new guy now probably believes the killing was justified.
I think of Trump's ongoing parade of press secretaries: The ones who have told the most stupid, hateful, ridiculous lies will be doing horrible things for years or decades in the future. Someday, documentaries might even talk about how sweet and good they were before they worked for Trump. That's how cognitive dissonance works.
Back to genAI
That's one reason, I think, why people I know who are otherwise very concerned with justice and their impact on larger systems--people who recycle, who buy fair-trade coffee, and who still mask in public--suddenly started telling me how awesome AI is and how the reports of its externalized harms are overblown. They used it for a while. If it was just a case of "this tool is very helpful," that's all they would say about AI, but they always go farther: now they believe its social, economic, political, and environmental harms must be minimal.
1 There are alternative theories (e.g., self-perception theory) that might explain the data just as well with more parsimony. They don't get much press outside academic psychology.
2 There is a bit more to the study, but I think this simplification keeps the core insights intact.
#uspol #cognitivedissonance #psychology #ai #genAI #llm