A bit of a scrape

Earlier today a friend/former colleague of mine linked to a dataset on the OpenScienceFramework with the commentary “OKCupid released some of their data!“  It turns out they didn’t release it, it was scraped.

The whole thing gives me a queasy feeling.  I’ve done and still do research on scraped data so believe me, I get the appeal, but I also see the drawbacks.  Science and technology by their very nature push ahead of the ability of law to regulate them.  We need to do better than “whatever, it’s legal” or we fall into moral hazard.  In this case, we have an ethical obligation to think carefully about digital privacy and not just assume that because we can access data, we _should_access it.  I’m not the only one who thinks so.

I’ll end this short post with an excerpt from an excellent commentary by the Council for Big Data, Ethics and Society which my friend madprime recommended to me:

It is no longer reasonable to claim that either the prior existence or the publicness of a dataset is a reasonable proxy for minimal informational risk posed by the data contained therein. We note that within the NPRM, public and private are used in a manner that leaves this regulatory gap open. The NPRM uses “public” to modify “datasets”—public describes the type of access to or availability of a dataset. In contrast, the NPRM uses “private” to modify “information” or “data”—private describes the expectation that a reasonable subject has about the relative availability of sensitive information. Publicly available datasets containing private data describes many of the sources most interesting to data researchers and practitioners, and are arguably most risky for subjects.