Last week, Facebook announced Graph Search, allowing users to discover more about their friends through intuitive, natural language queries. Though the engineering behind Graph Search is exciting and the immediate utility it provides users is apparent, its arrival necessitates a consideration of the motives behind this feature release.

Graph Search was actually around years ago in the form of advanced filters, but the fact that Facebook now considers this the “third pillar” (with Newsfeed and Timeline the other two) suggests that the most recent iteration of search is driven by a few strategic plays:

1. Improve the process of finding new friends
2. Increase advertising revenue through search
3. Incentivize and formalize the purpose of sharing

Most Facebook users add friends only after they’ve met them in person. Though this proportion might not significantly change in the post-Graph Search world, Graph Search reduces the friction of finding others with similar interests and tastes, thereby encouraging additional Facebook friendships. Monetizing the “personalized search” queries that users enter gives Facebook the opportunity to harness both intent and personal information. Perhaps most importantly, Graph Search finally offers a tangible user scenario where sharing and likes serve some sort of higher purpose.

As it stands, Facebook likes are incredibly noisy. No one really knows what a “like” means; some might use it ironically, while others use it to enter raffles offered by Facebook pages. Ideally Graph Search gives meaning to the “like”, re-orienting this action to one used exclusively for curating a personal profile. However, I’m not convinced that Graph Search will help fix the very signals that it depends upon.

I understand the idea that seeing helpful results might incentivize users to share additional information, but this only works if the link between sharing and helpful results is quite clear. Consider contributors in other online communities, like Yelp, Wikipedia, Quora, or StackOverflow: all rely upon a small group of key contributors who are driven by an intrinsic motivation to create content. In order to compel those who are not intrinsically motivated to contribute, one would need to clearly demonstrate the benefits of contribution. In the case of Graph Search, these benefits are not readily apparent. Graph Search alone will not change the way users like or share on Facebook.

Moreover, unlike the aforementioned communities which publicly feature content, Facebook can only display content in a manner that obeys existing privacy settings; it needs much more than a small community of dedicated power users to ensure that everyone is delivered meaningful content. If a large proportion of users aren’t compelled to curate their Facebook profiles with their preferences and interests, then Facebook fails to achieve all three of its strategic goals.

Just like Newsfeed and Timeline, Graph Search strives to make it easier for users to find information, redefining privacy expectations on Facebook. There is such thing as “security through obscurity,” and everyone takes this for granted in daily life. danah boyd gives a particularly apt analogy:

Think about a cafe that you like to visit. This is fundamentally a public space. There’s a possibility that you’ll intersect with all sorts of different people, but there are some people who you believe you are more likely to interact with than others. You have learned that you’re more likely to run into your neighbors and you’d be startled if your mother “popped in” since she lives 3000 miles away. You may have even chosen this particular cafe in the hopes of running into that hottie who you have a crush on or avoiding your ex who lives in a different part of town. You have also come to understand that physics means that there’s a limit on how many people will be in the cafe. Plus, you’d go completely bonkers if, all of a sudden, everyone from your childhood magically appeared at the cafe simultaneously. One coincidence is destabilizing enough; we can’t really handle a collapse in the time-space continuum.

When people assess a situation, they develop mental models based on probability calculations and the expectations they bring to the table. They make guesses about who is more or less likely to run across them. Their calculations are completely reasonable, as it’s an efficient way of getting a decent handle on the social context, even if they are sometimes wrong. This is true both offline and online. People need to know how to behave so they use whatever information is available to them to make their best guess.

Graph Search will force some to go back and review all of their likes, while others may be left with the unfortunate consequences of leaving embarassing content that appears in search results. Without careful curation of the Activity Log, Facebook offers no other protection (there is no opt-out of Graph Search). To deny users this basic control is unwarranted.

Each major Facebook update implicates some sort of privacy tradeoff, and unsurprisingly, the rhetoric surrounding these discussions is highly reminiscent of debates over national security surveillance. Each time we are told that sacrificing a bit of privacy for the greater good is worth it. After all, we’re only making it easier for the inevitable reality that our information may be accessed. Though we theoretically have safeguards intended to protect our privacy, in practice, they’re insufficient. Just as the executive branch dodges oversight, Facebook’s frequent and opaque privacy changes leave users unaware and confused about who has access to their content.

Facebook’s mission is to make the world “more open and connected.” This is advantageous to Facebook as a business, but the idea that radical transparency benefits human society is weak at best. There’s a discussion to be had about how the architecture of Facebook impacts human psychology and interaction, but on the axis of privacy alone, Graph Search represents an unprecedented shift, on par with the arrival of the Newsfeed years ago. Yet unlike the Newsfeed, it’s clear that neither Facebook nor its users are quite prepared for Graph Search.