Lesson: Open Government with Aneesh Choprah

Step #8 Security: Privacy and security are risks associated with open government

Privacy and security, so what we often are concerned about is what's referred to as the Mosaic Effect. That is any given database might feel as if its release to the public is accessible, okay, and won't violate your privacy. But the concern is if someone combined dataset one, two and three, perhaps with a private dataset, it would unexpectantly reveal a private citizen's information.

The Mosaic Effect has been a big concern and thus over time, you have seen the data.gov platform constrain or restrict the datasets that it publishes for those that are at risk of potentially contributing to the Mosaic Effect. So the biggest concern is privacy and security, though generally speaking, if the data would've been made available anyway, if you requested it through a Freedom of Information Request, then we felt that the default should be you should put it on the data.gov. These are the areas of greatest concerns.

It is definitely the case that if you have open data, people might consume that data to harm and the reality is we have to accept a certain level of risk by having opened up the data, not because we are willing to tolerate risks, but rather the information would have been made available anyway. So if I wanted a database of all installations that might have some strategic value, much of that information is already available.

We made these judgments on privacy and security as to whether or not we will be compromising the individual security of the American people and we did take a filter to make sure that the Security Council and others said, "Look, you can or you can't release this particular dataset." But by and large, you cannot control what people do with the data, that's the whole point of open, and so our hope, thus far, is more good than bad and the bad, largely a clumsy implementation. But it is conceivable that people might use this information inappropriately.