Health Matrix·Volume 24·2014
Big Data Proxies and Health Privacy Exceptionalism
86
networked devices.
119
Some of this data may still be unused by big
data because it is “dark data” that has been left over or discarded
from other processes and not yet leveraged,
120
or, in the words of
Andrew McAfee and Erik Brynjolfsson, “[T]here’s a huge amount of
signal in the noise, simply waiting to be released.”
121
Consider just one example of a recognized big data source: social
media interactions. Michal Kosinski and colleagues analyzed the
Facebook “likes” of almost 60,000 volunteers. Using big data tech-
niques the researchers were able to predict “sexual orientation,
ethnicity, religious and political views, personality traits, intelligence,
happiness, use of addictive substances, parental separation, age, and
gender” and speculated that “given appropriate training data, it may
be possible to reveal other attributes as well.”
122
As hypothesized by
FTC Commissioner Julie Brill:
[W]e can easily imagine a company that could develop algo-
rithms that will predict . . . health conditions – diabetes, cancer,
mental illness – based on information about routine transactions
http://www.nature.com/srep/2013/130325/srep01376/pdf/srep01376.
pdf.
117. See, e.g., Emily Steel & April Dembosky, Health App Users Have New
Symptom to Fear, F
IN. TIMES (Sept. 1, 2013),
http://www.ft.com/intl/cms/s/0/97161928-12dd-11e3-a05e-
00144feabdc0.html.
118. See, e.g., Amy Dockser Marcus & Christopher Weaver, Heart Gadgets
Test Privacy-Law Limits, WALL ST. J. (Nov. 28, 2012),
http://online.wsj.com/article/SB10001424052970203937004578078820874
744076.html.
119. See, e.g., Evgeny Morozov, Requiem for Our Wonderfully Inefficient
World, SLATE (Apr. 26, 2013),
http://www.slate.com/articles/technology/future_tense/2013/04/senor
_based_dynamic_pricing_may_be_efficient_but_it_could_create_in
equality.html.
120. Isaac Sacolick, Dark Data – A Business Definition, S
OCIAL, AGILE, AND
TRANSFORMATION (Apr. 10, 2013),
http://blogs.starcio.com/2013/04/dark-data-business-definition.html
(“Dark data is data and content that exists and is stored, but is not
leveraged and analyzed for intelligence or used in forward looking
decisions.”).
121. McAfee & Brynjolfsson, supra note 76, at 63.
122. Michal Kosinski et al., Private Traits and Attributes Are Predictable
from Digital Records of Human Behavior, 110 P
ROCEEDINGS OF THE
NAT’L ACAD. SCI. 5802, 5805 (2013). See also Katie Lobosco, Facebook
Friends Could Change Your Credit Score, CNN (Aug. 27, 2013),
http://money.cnn.com/2013/08/26/technology/social/facebook-credit-
score/index.html?hpt=hp_t2 (describing how “some financial lending
companies have found that social connections can be a good indicator of
a person’s creditworthiness”).