Boing Boing Staging

Crystal, an app that attempts to summarize your personality

obama on crystal
Crystal is an app that attempts to summarize your personality by analyzing your online presence. My friend the philosopher Evan Selinger wrote a smart assessment of the problems with this approach — and then ran a Crystal search on me, with somewhat hilarious results.


The theory behind Crystal is that it’ll help improve communication online. Before you contact someone for the first time, you can generate a Crystal profile to understand how they prefer to communicate — then “tweak your prose to conform with a recipient’s conception of an ideal note,” as Selinger notes.

Does it work? At The Next Web, they found it “often eerily accurate”, while at The Daily Dot, Selena Larson called it “creepily accurate”.

Selinger didn’t find it quite so impressive. He ran a test on his future boss, Jules Polonetsky, but the results were, to Polonetsky’s eyes, “very wrong in its basic efforts to assess me. Unless, of course, the program knows me better than I know myself.” Selinger also generated a Crystal profile of Obama, which found that the president is “occasionally rambling in conversation”. Whether or not one sees this is true depends on context — what you regard as longwindedness, I may regard as thoughtfulness, Selinger points out. This is precisely why it’s hard to create an objective, one-size-fits-all measurement of someone’s human spirit.

I DMed Selinger and asked him to run a Crystal search on me. These were the results (pardon the my lame photoshop cropping):

As with Polonetsky, I found this didn’t ring true. Sure, I think I have good EQ, and rely on itevery day. But the idea that I trust “feelings and gut instinct more than rules or logic”? That seems nuts. I frequently report on science, a realm where “going with your gut” doesn’t get you very far. And as for not trusting rules, my friends and family routinely complain about precisely the opposite — I have a nearly Soviet level of obsequiousness when faced with official procedure. (Which, being both Canadian and former Boy Scout, one might expect.)

As for the other advice about approaching me? Sure, some of it seems perfectly fine — “When emailing Clive, don’t use a formal greeting and closing” — though, as with a horoscope, it’s also sufficiently open-ended as to apply to a large percentage of people. (“Tuesday, you will face many challenges.”) In fairness to Crystal, they admit that their accuracy confidence is only 47%, since they found “limited data” on me. The limited data probably stems from the fact that I don’t use LinkedIn at all, Facebook rarely, and for whatever reasons they didn’t find the 2,000-or-so blog posts I’ve written to be useful.

For Selinger, though, the problem with Crystal isn’t in its current manifestation. The question is what happens if it — and personality-analyzing apps of its ilk — become good enough that people really start to rely on them. As he writes:

Chiefly, there’s the problem of distortion. If people don’t realize that a profile is inaccurate, they can be nudged to come to false conclusions. Imagine technology such as Crystal creating biased characterizations that influence how people view politicians or job candidates. It’s not hard to see how discrimination and prejudice can be inflamed in those cases. [snip]

If Crystal is just the beginning of a new category of similar apps that rely on algorithms to tell people what to say and how to write, these technologies will not only have a troubling effect on privacy but also begin to strip the character and individuality from communication in favor of banal, machine-generated prose.

Over at Slate, Elana Zeide noted another issue: Even if we know an algorithmic assessment is sketchy, we can still be influenced by it:

Confirmation bias is strong, however, and the program primes users to look for predicted attributes. Crystal predicted that it doesn’t come naturally to an old friend of mine to speak with her hands. I was suspicious of this inference because she has a minimal social media presence and few, if any, photos online. The program was inaccurate, but I found myself scanning memories several times for instances supporting that inference. Even knowing Crystal’s accuracy rating was only 49 percent, I still doubted my own perception, which was formed by more than 20 years of firsthand data.

Then again, you’d expect me to be a bit skeptical about Crystal, wouldn’t you? After all, according to their own assessement, “It does not come naturally to Clive to stay loyal to trusted companies ands brands”, heh.

Exit mobile version