Privacy

  • Mass Spying Isn't Just Intrusive—It's Ineffective. Of course it doesn't work. But affected actors (i.e. three letter agencies) can then claim they are doing something and they need more resources -- to do more, not to do it better.

    But the track record of the collection programs Edward Snowden revealed provides little evidence that massive surveillance will help us identify future terrorist attacks or mitigate these new risks. American spies’ allegiance to massive surveillance is based on faith, not track record. The Boston Marathon bombing in April of 2013 illustrates how broad proactive surveillance is no panacea against attacks. The NSA was conducting its massive spying at the time, and the attacks happened anyway.

  • New Bill Would Force NYPD to Disclose Its Surveillance Tech Playbook.

    The bill, named the Public Oversight of Surveillance Technology (POST) act, would require the NYPD to detail how, when, and with what authority it uses technologies like Stingray devices, which can monitor and interfere with the cellular communications of an entire crowd at once. Specifically, the department would have to publicize the "rules, processes and guidelines issued by the department regulating access to or use of such surveillance technology as well as any prohibitions or restrictions on use, including whether the department obtains a court authorization for each use of a surveillance technology, and what specific type of court authorization is sought."

Tech

  • Palantir Provides the Engine for Donald Trump's Deportation Machine.

    Immigration and Customs Enforcement is deploying a new intelligence system called Investigative Case Management (ICM), created by Palantir Technologies, that will assist in President Donald Trump's efforts to deport millions of immigrants from the United States.

  • Facebook's New Algorithm Combs Posts to Identify Potentially Suicidal Users.

    This week, Facebook announced beefed up suicide prevention tools that use algorithms to scan posts and look for potentially suicidal language—posts about sadness, pain, and not being able to take it any more, for example—as well as take note of comments on posts that may signal an issue—things like "are you okay?" and "I'm here for you."

  • How Google Street View Images Reveal the Demographic Makeup of the US. Link to PDF paper here.

    Here, we present a method that determines socioeconomic trends from 50 million images of street scenes, gathered in 200 American cities by Google Street View cars. Using deep learning-based computer vision techniques, we determined the make, model, and year of all motor vehicles encountered in particular neighborhoods. Data from this census of motor vehicles, which enumerated 22M automobiles in total (8% of all automobiles in the US), was used to accurately estimate income, race, education, and voting patterns, with single-precinct resolution. (The average US precinct contains approximately 1000 people.) The resulting associations are surprisingly simple and powerful. For instance, if the number of sedans encountered during a 15-minute drive through a city is higher than the number of pickup trucks, the city is likely to vote for a Democrat during the next Presidential election (88% chance); otherwise, it is likely to vote Republican (82%).

  • How Uber Deceives the Authorities Worldwide.

    Uber has for years engaged in a worldwide program to deceive the authorities in markets where its low-cost ride-hailing service was resisted by law enforcement or, in some instances, had been banned.

    The program, involving a tool called Greyball, uses data collected from the Uber app and other techniques to identify and circumvent officials who were trying to clamp down on the ride-hailing service.

  • Google's anti-trolling AI can be defeated by typos, researchers find. Seems like we are at the beginning of the spam wars again, but with trolls instead of Viagra.

    In a paper published on February 27, Hossein Hosseini, Sreeram Kannan, Baosen Zhang, and Radha Poovendran demonstrated that they could fool the Perspective AI into giving a low toxicity score to comments that it would otherwise flag by simply misspelling key hot-button words (such as "iidiot") or inserting punctuation into the word ("i.diot" or "i d i o t," for example).

Visualizations


Data Links is a periodic blog post published on Sundays (specific time may vary) which contains interesting links about data science, machine learning and related topics. You can subscribe to it using the general blog RSS feed or this one, which only contains these articles, if you are not interested in other things I might publish.

Have you read an article you liked and would you like to suggest it for the next issue? Just contact me!