Stories

Suit to let researchers break website rules wins a round

Hand reaches out from data
Thomas M. Scheer / EyeEm via Getty Images

Anyone following Facebook’s recent woes with Cambridge Analytica might be surprised to hear that there's a civil liberties argument for swiping data from websites, even while violating their terms of service. In fact, there's a whole world of situations where that thinking could apply: bona fide academic research.

What's new: On Friday, a judge in a D.C. federal court ruled that an American Civil Liberties Union-backed case trying to guarantee researchers the ability to break sites' rules without being arrested could move forward, denying a federal motion to dismiss.

"What we’re talking about here is research in the public interest, finding out if there is discrimination,” Esha Bhandari, an ACLU attorney representing the academics, told Axios.

The details: A handful of researchers and First Look Media (which operates The Intercept and other sites) would like to use bots and create dummy accounts to test the behavior of employment and real estate websites.

  • The researchers are studying whether machine-learning algorithms on employment and real estate websites might have developed gender or racial bias. To do that, they would set up multiple similar accounts, changing only minority or gender status between them, and apply for jobs or housing.
  • That might violate the sites’ terms of service — and doing so, some courts have ruled, constitutes a violation of the Computer Fraud and Abuse Act (CFAA), the major U.S. anti-hacking law.

Why it matters: Knowing whether or not websites are biased against women and minorities is a public good. But sites aren’t always eager to help researchers reach those kinds of conclusions about them. The Department of Justice is not currently using the CFAA to bludgeon researchers who expose discrimination online. But that’s only by internal policy. Without courts clarifying the law (or legislators changing it), that threat could hang over researchers and their work.

The decision: Judge John D. Bates denied the Justice Department’s argument that the researchers had no guaranteed legal right to access data on a publicly visible website.

Why this outrages people less than Cambridge Analytica: The results of the study will be for public consumption. And — obviously — no data will be used to target ads.

What the case doesn’t decide: The CFAA is not the only impediment to this kind of research.  Many academic publishers bar studies that depend on data accessed in violation of terms of service. Without a place to publish, there’s often no incentive to do the research.

  • Terms of service can defend sites' intellectual property and user privacy, but can also be used to protect a site's reputation and hide misbehavior. The CFAA allows for civil lawsuits, and a company could sue researchers producing research it doesn't like. The civil aspects of the law are not part of this case.