The shift of library services to online interfaces has led to an explosion in the potential for data gathering, and also to a growing conversation about how the data should and could be used. This past year has witnessed a strong dialogue about libraries’ responsibility for maintaining the privacy and security of the data. Leading experts have pointed out the astonishing number of ways that privacy and security are unintentionally compromised in libraries’ everyday service environment. Protecting the privacy of library users has long been seen as a core value of the profession, and the relevance of this value is regularly defended as paramount. But I worry that we have been framing our discussion about privacy as library-centric and insufficiently user-centric.

The library-centric approach to privacy frames it as largely a black-or-white proposition: libraries can offer services that may add value but equally may pose concerns. In this conception, the library has a choice, whether to offer or support such services, or to avoid doing so.

But a user-centric approach to privacy would recognize that users have certain needs, or desire certain types of services or experiences. Many providers, including the library, are competing with one another to serve these needs. If only advertising-driven commercial providers are in the marketplace to provide these services, then users will turn to them, in spite of their poor protection of the privacy imperative.

This latter approach is more humble, recognizing the limitations of the library’s agency. If the library chooses not to provide a given service out of fear of violating user privacy, this does not necessarily protect the users’ privacy. Rather, in the case of highly valued services, it leads the user to providers that are potentially far more problematic, and over which the library, and the academy more broadly, exercises little if any control or influence. For example, libraries have studiously avoided gathering data that could personalize discovery, ceding this terrain to a variety of other parties such as Elsevier and Google. With the best of motives, librarians have ignored emerging community needs.

It is too simple for the library to avoid providing a given service simply because there is a risk of some privacy concern. A user-centric model frames the library’s choices with more nuance and looks to find the right balance. In my view, where the community places value in a service that requires personal data, the library should build a service model that safeguards user privacy more so than would the alternative.