The way privacy is encoded into software doesn’t match the way we handle it in real life.
Each time Facebook’s privacy settings change or a technology makes personal information available to new audiences, people scream foul. Each time, their cries seem to fall on deaf ears.
The reason for this disconnect is that in a computational world, privacy is often implemented through access control. Yet privacy is not simply about controlling access. It’s about understanding a social context, having a sense of how our information is passed around by others, and sharing accordingly. As social media mature, we must rethink how we encode privacy into our systems.
Privacy is not in opposition to speaking in public. We speak privately in public all the time. Sitting in a restaurant, we have intimate conversations knowing that the waitress may overhear. We count on what Erving Goffman called “civil inattention”: people will politely ignore us, and even if they listen they won’t join in, because doing so violates social norms. Of course, if a close friend sits at the neighboring table, everything changes. Whether an environment is public or not is beside the point. It’s the situation that matters.
Whenever we speak in face-to-face settings, we modify our communication on the basis of cues like who’s present and how far our voices carry. We negotiate privacy explicitly—“Please don’t tell anyone”—or through tacit understanding. Sometimes, this fails. A friend might gossip behind our back or fail to understand what we thought was implied. Such incidents make us question our interpretation of the situation or the trustworthiness of the friend.
All this also applies online, but with additional complications. Digital walls do almost have ears; they listen, record, and share our messages. Before we can communicate appropriately in a social environment like Facebook or Twitter, we must develop a sense for how and what people share.
Danah Boyd is a social-media researcher at Microsoft Research New England, a fellow at Harvard University’s Berkman Center for Internet and Society, and a member of the 2010 TR35.