Presenter Predicts 2020 Census Manipulation

Drew Kaplan ’20, Associate Opinion Editor

danah boyd predicted at the Fairness and Accountability in Algorithmic Culture that the 2020 census would be manipulated because of public distrust.

Sponsored by the math and computer science departments, boyd, who styles her name in the lower case (“to reflect [her] mother’s original balancing and to satisfy [her] own political irritation at the importance of capitalization,” according to her Wikipedia page,) opened with the assertion that technology has increasingly become bound to us, and that we are being exposed to greater amounts of technology at younger ages. However, the data generated from this increased use of technology is now stored and sold to businesses. boyd told a story about how, in 2012, Target had sent an ad targeted toward expecting mothers to a 16-year-old girl; Target had figured out that she was pregnant before she had told her family. “Nothing is more valuable than identifying someone who as at a significant life change point,” said boyd. She said there is a whole system in place to identify when life events occur, and that it is possible to cross reference a date of marriage with a zip code to predict when a child will be born.

boyd explained how the perception of big data has changed in recent years. “Big data is a fuzzy concept” that has become “very popular in a very short period of time.” An issue is that, amongst the hype, people have tried to gather more and more data, with the hope that somehow computers will resolve everything. boyd said that while this has improved some things, medical discoveries have been made by computers that would be unnoticeable to human doctors, and are making artificial intelligence possible. Public opinion around big data has “turned into a narrative of big brother;” the level of surveillance surrounding the mass collection of data makes people uncomfortable.

However, there are flaws in how the data is collected and used. “We see all sorts of cultural issues work their way into data systems” said boyd. “There is nothing neutral about big data.” Because “most of the data we’re contending with is produced for a number of reasons,”

Biases are either expressed in how the data is collected or the results found from the data. Part of the issue is that not all data is collected in the same circumstances. Some of it is surrendered voluntarily, contracts we sign or polls we take part in, or under duress, the collection of DNA by police during a traffic [stop]. The DNA can then be sent through all layers of government, and potentially used against the person later.

However, the largest flaw is that people will intentionally try to mess with data. People have tricked self-driving cars into misidentifying road signs, and the redefinition of the surname of Rick Santorum, which boyd declined to define for the audience. boyd also predicted that there will be a concerted effort to manipulate the upcoming census by associating the proposed “Muslim ban” with Japanese internment during World War II. If people fear that their census data will be used against them, they are less likely to answer honestly, which skews the results. “We’re going to see a collapse of the 2020 census” she said.

boyd finished by saying that the state of things to come is hazy. “We often have the political will to identify a problem, but not to fix them,” boyd said, “We’re exhausting with having to deal with breach after breach.” boyd added that more accountability is needed in the system, though introducing it is difficult. She said that “Governments have abused every new technology,” and that she is “very grateful to [Snowden]” for his role in exposing how our data was being used by the government.

Responses to the talk were positive. Before the talk began, Ashir Borah ’19, said “I think it’s a very interesting talk, because being fair, and still being able to achieve optimality can be a very interesting ethical topic.”

Noah Hunt-Issak ’20, said “I thought the speaker was extremely knowledgeable. I’ve always thought the idea of machine learning and AI was really cool, especially because we can create machines which notice patterns and “form conclusions” in a way humans can’t.”

Erik Love, Assistant Professor of Sociology, said “I’ve been a fan of danah boyd’s work for a long time. A lot of what she had to say tonight was really insightful.”