The Human Psychome
From the 1930s through the 1950s, corporations used personality tests to analyze employment suitability and match candidates with jobs, most notably the Thematic Apperception Test (TAT, 1930s), the Minnesota Multiphasic Personality Inventory (MMPI,1943), and Myers-Briggs (1944). Many of us took abbreviated forms of Myers-Briggs as a game among friends–we enjoyed seeing whether the test’s four-letter characterization matched what others really thought of us. These “psychometric” assessments were sufficiently alarming for William Whyte in The Organization Man (1956) to include an appendix, “How to Cheat on Personality Tests.” If you followed his advice, you would not be summarily labeled psychotic or homosexual (they cared about that then…) based on what you thought of a picture or whether you went to a museum on Sunday.
The personality-test fad abated, and it’s hard to imagine any children of the 60s or 90s subjecting themselves to this kind of psycho-pseudo-babble analysis—the basis is just too thin, and we are each our own person. Or so we think. In the end, the conforming and conformist corporate practices of that era went extinct, and their replacements celebrate eclecticism rather than run from it. Or so we hope.
But, these tests are back—in spades—largely because we now believe that big data, big sensing, big monitoring, and big analysis can catalogue and quantify everything. Watson and Crick defined the structure of DNA; Ventner et al defined a whole person’s DNA structure: the human genome; we’re creating a human biome for the DNA in our body that is not in our own cells. What we thought was a good joke as teenagers is not only taken seriously now, but is being used to create the psychological equivalent of the human genome or biome: a human psychome. If we know enough about what you say and do, your psychome will seemingly predict your response to a sales pitch, propaganda, or any stimulus.
If we know enough about what you say and do, your psychome will seemingly predict your response to a sales pitch, propaganda, or any stimulus.
The TAT and the MMPI tests were taken deliberately. Now, we take such tests behaviorally, every time we purchase something in a store or on the Internet, in the act of a purchase rather than as a questionnaire. The results? They’re kept siloed, out of view. You don’t even know the grading. But more agonizing—or creepy, as is the common characterization in the media—is the fact that these behavioral computations are used to second-guess what you will do, buy, and think. Who likes to be second-guessed?
Will this work? Well, if you believe Ray Kurzweil and think that in short order you will be able to download your brain, then clearly a corollary of that is analyzing your brain. But if you think that the more we learn about the mind the more we learn we have to learn, then you should be very suspicious about both the claims and the applications of those would claim to know what you think.
I find solace in the work of Kahneman, Tversky, and Ariely. They took some subtle aspects of human nature and catalogued and measured them. While their work in behavioral science can be a textbook for propagandists, it is also a defense, and it teaches us that when it comes to how we think and act, the cards are not all in.
It turns out that privacy is not about what you disclose when you know you are doing it. If you use your credit card in a restaurant, you understand they know your name. My local restaurant knows me as Andy because I give them my name and phone number when I order takeout. Each time I order, they say, OK, Andy. That’s creepy, but not scary. Scary is when they assume that pork-eaters favor gun rights and add a cookie paid for by the gun lobby. In other words, it is not the lack of privacy alone that alarms; rightly or wrongly, we are inured to it. Rather, it is the use to which your identity is put that should alarm us more. It’s the computation that is the greater invasion of privacy. Deducing your taste is as invasive as revealing your medical records.
Deducing your taste is as invasive as revealing your medical records.
And neither your acts nor your medical records paint a full picture. In genetics we know there are some traits that are a mix of genetic predisposition and environment. Genes don’t define the whole person and can’t predict them. Neither do our observable actions. Even if there is a psychome to find, it is unlikely to be our whole picture, nor be predictive in any given situation. Maybe that’s why these hamfisted attempts are so scary — are they talking to me or some cardboard effigy?
By some measure, it is bad that private corporations or governments are building my psychome and worse that their structure is built on sand. To the extent that you are your psychome, they are stealing a part of you. And what they know is no more accurate than what your spouse, friend, or partner knows — and perhaps a good deal less so. If we have a psychome at all, let’s at least build a world where we can each own ours. And where we use it to match our behavior with our aspirations, not just for sales.
Andy Lippman is associate director of the Media Lab, and the head of the Viral Communications research group.