In my previous post about Cambridge Analytica, I wrote about how those incensed that manipulative algorithms had swayed susceptible voters towards Trump and Brexit were ignoring the structural issues that had made the voters susceptible in the first place. Over the weekend, I read a piece which said psychographic targeting is nowhere near as effective as it was being made out to be and that Cambridge Analytica had a pretty spotty record of success. Even before I read the article, I had become skeptical of the whistleblower Christopher Wylie after I watched his testimony to a British Parliamentary committee. When asked if Cambridge Analytica's work on Vote Leave could have affected the results of the referendum, he replied that when an athlete dopes he is banned outright, without splitting hairs over how much had he doped, etc. He was before the committee as an expert witness and not to wax on with his political views or what should be done next. The more I saw of him and the Guardian's fawning coverage of him, the more he struck me as either a grandstander hoping to stretch his five minutes of fame or the protagonist of a plot to force a second referendum.
I know too little about technology or psychology to have an informed opinion on the matter, and I admit I initially believed that Cambridge Analytica had succeeded in manipulating the vote. When we don't understand something, we bestow it with great powers in our minds. But, I think this episode raises an interesting question about the primacy of structure or agency that has befuddled social scientists for ages. Do a person's political views depend more on her personality or her circumstances, on her psychology or her sociology? If you think it depends mostly on psychology than you would tend to agree that psychological manipulation can be quite effective. If you think it depends more on her social circumstances and life experiences, you would be much more skeptical of this claim.
P.S. I have been doing some more critical reading on Cambridge Analytica and the upshot is this: It is difficult to predict personality traits from facebook behaviour, it is even more difficult to estimate political orientation based on personality traits, and it is insanely difficult to change political orientations, which is what Cambridge Analytica is accused of pulling off.
I know too little about technology or psychology to have an informed opinion on the matter, and I admit I initially believed that Cambridge Analytica had succeeded in manipulating the vote. When we don't understand something, we bestow it with great powers in our minds. But, I think this episode raises an interesting question about the primacy of structure or agency that has befuddled social scientists for ages. Do a person's political views depend more on her personality or her circumstances, on her psychology or her sociology? If you think it depends mostly on psychology than you would tend to agree that psychological manipulation can be quite effective. If you think it depends more on her social circumstances and life experiences, you would be much more skeptical of this claim.
P.S. I have been doing some more critical reading on Cambridge Analytica and the upshot is this: It is difficult to predict personality traits from facebook behaviour, it is even more difficult to estimate political orientation based on personality traits, and it is insanely difficult to change political orientations, which is what Cambridge Analytica is accused of pulling off.
No comments:
Post a Comment