https://theintercept.com/2016/04/14/in-undisclosed-cia-investments-social-media-mining-looms-large/
The CIA Is Investing in Firms That Mine Your Tweets and Instagram Photos
Lee Fang
April 14 2016, 1:57 p.m.
Bruce Lund, a senior member of In-Q-Tel’s technical staff, noted in a 2012 paper that “monitoring social media” is increasingly essential for government agencies seeking to keep track of “erupting political movements, crises, epidemics, and disasters, not to mention general global trends.”
The recent wave of investments in social media-related companies suggests the CIA has accelerated the drive to make collection of user-generated online data a priority. Alongside its investments in start-ups, In-Q-Tel has also developed a special technology laboratory in Silicon Valley, called Lab41, to provide tools for the intelligence community to connect the dots in large sets of data.
In February, Lab41 published an article exploring the ways in which a Twitter user’s location could be predicted with a degree of certainty through the location of the user’s friends. On Github, an open source website for developers, Lab41 currently has a project to ascertain the “feasibility of using architectures such as Convolutional and Recurrent Neural Networks to classify the positive, negative, or neutral sentiment of Twitter messages towards a specific topic.”
Collecting intelligence on foreign adversaries has potential benefits for counterterrorism, but such CIA-supported surveillance technology is also used for domestic law enforcement and by the private sector to spy on activist groups.
Palantir, one of In-Q-Tel’s earliest investments in the social media analytics realm, was exposed in 2011 by the hacker group LulzSec to be in negotiation for a proposal to track labor union activists and other critics of the U.S. Chamber of Commerce, the largest business lobbying group in Washington. The company, now celebrated as a “tech unicorn” — a term for start-ups that reach over $1 billion in valuation — distanced itself from the plan after it was exposed in a cache of leaked emails from the now-defunct firm HBGary Federal.
Cover of the document obtained by The Intercept.
Yet other In-Q-Tel-backed companies are now openly embracing the practice. Geofeedia, for instance, promotes its research into Greenpeace activists, student demonstrations, minimum wage advocates, and other political movements. Police departments in Oakland, Chicago, Detroit, and other major municipalities have contracted with Geofeedia, as well as private firms such as the Mall of America and McDonald’s.
Lee Guthman, an executive at Geofeedia, told reporter John Knefel that his company could predict the potential for violence at Black Lives Matter protests just by using the location and sentiment of tweets. Guthman said the technology could gauge sentiment by attaching “positive and negative points” to certain phrases, while measuring “proximity of words to certain words.”
Privacy advocates, however, have expressed concern about these sorts of automated judgments.
“When you have private companies deciding which algorithms get you a so-called threat score, or make you a person of interest, there’s obviously room for targeting people based on viewpoints or even unlawfully targeting people based on race or religion,” said Lee Rowland, a senior staff attorney with the American Civil Liberties Union.
She added that there is a dangerous trend toward government relying on tech companies to “build massive dossiers on people” using “nothing but constitutionally protected speech.”