Saving Facebook: Yale ISP Lunch Speaker Series with James Grimmelmann
by ellanso | April 15, 2009 | news and ideas | Comments Off
Today’s ISP Lunch Speaker is James Grimmelmann of New York Law School, who is presenting a talk entitled “Saving Facebook”, based on his forthcoming article in the Iowa Law Review (Vol. 94). Slides from this presentation are available at his website.
What is it about technology that causes people to do really dangerous things? And how should we respond to people using technology in foolhardy ways?
Ghost Riding the Whip Association Facebook page — presenting the dangerous things people do, associated with their real names, to the world at large. What responses should we have?
definitions
Social Network Sites: The emphasis is on the network of connections among people, rather than the act of networking.
Profiles create an individual’s online identity, links provide connections based on relationships between and among people. The social graph makes the community of people, defined by these links, easy to perceive.
Identities connected by relationships to form a community. I — you — them.
Examples: Friendster, Facebook, MySpace are all explicitly social network sites. YouTube, Flickr, and Twitter have lots of social network features.
social motivations
Why do people go to these sites?
Identity – tell people who you are, craft a profile as a tool of impression management (pick best photo of yourself for your profile, name the right bands (or the appropriately ironically un-hip bands). The messages on your Wall, the comments you leave — all of this helps create the sense of who you are.
Relationship – making new friends, keeping current with what people you already have relationship with are doing, tracing the connections among your friends
Community – social network effects, everyone in a high school joining Facebook because that’s where the social life of the school is, mass migrations of groups from one social network platform to another, visualization of how the people in your life are connected, and conspicuous/competitive friending. Being friends with the right people. These are all deeply wired human social impulses, just taking place in a different environment.
social risk assessment
But people are bad at measuring the privacy risks of these behaviors. We think of safety in numbers but this can lead to detrimental cascade effects, where enough people engage in a risky behavior that it seems like a non-risky behavior, even though no one’s done a rational analysis of the privacy risks or concerns.
The interface of the sites can lead people into a false sense of security: posting a message on someone’s Wall seems like a face-to-face encounter (you see their face, it’s a one-to-one conversation), but it’s actually viewable to everyone in the network. Can also feel lulled into a sense that you know the person better than you actually do.
People are operating with different privacy expectations, and the digital medium sheds a lot fo the context clues that would allow people to know that “I didn’t mean for you to forward that email to all of our friends!!”
Mutual surveillance helps to enforce privacy guarantees and etiquette, but the social cuing about this is not as easily perceived in the digital medium. (No sidelong glances or staring people down if they’re watching you in a way that makes
social privacy harms
Disclosure: of information that you might not want to disclose to everyone with an internet connections. Examples: employers seeing your interests that include “smokin’ blunts”, breaking up with people via Facebook status messages
Surveillance: it’s creepy to know that someone’s watching. There was an uproar when Facebook first introduced the News Feed, a real-time mass surveillance tool. Ultimately, people adjusted to the new feature, but not before demanding opt-out options.
Instability: you can’t depend on the platform continuing to work in the same way. Facebook profiles used to be non-searchable, but now you have to opt-out of searchability. Facebook Beacon allows your other web use to be published in your newsfeed, leaking information that Facebook users couldn’t have anticipated.
Disagreement: Social network sites bring a lot of interactions to light that might have otherwise never be addressed. Receiving friend requests from people you don’t want to friend: is it rude to ignore them? MySpace’s Top Eight friends feature causes a lot of angst among the teenage set. Also, de-tagging incriminating and/or personal photos that you don’t want to have linked to your Facebook profile (though your friends who are tagging you want to associate this info with you).
Spillovers: Your choices affect my visibility – if you set your viewability to “friends of friends” and have a promiscuous friender as a friend, that’s a lot of people who can view your profile. Inference on social network graphs: how many gay friends do you need to have before others can make inferences about your sexuality? Or friends of a particular age range and nationality before someone can infer your age and nationality?
Denigration: Other users disrupt your ability to present yourself as you want to be seen. If you have lame friends that say lame things on your Wall, you’re probably pretty lame. Worlds collide on these platforms, because in real life you have multiple disconnected social networks, but they converge at the node that your profile represents.
solutions
Things that Don’t Work:
Privacy Policies – people don’t read them, and if they do, they don’t understand them. Also, privacy policies tend not to promise much in the way of privacy.
Technical Controls – Facebook’s controls are outstanding: they’re powerful, simple-to-use, offer fine-grained control, but people don’t use them. Further, technical rules don’t solve social problems. We have ambiguous relationships in real life (where you’re not really sure about your friend status with someone), and having multiple technical means to define friendships doesn’t help when you don’t have a definition in the first place. They also can’t make your friends more trustworthy.
Data Ownership – Being able to take your data with you is handy, to be sure (keeps you from being chained to a single platform that develops increasingly bad policies), but your data’s not just about you, it’s about your friends, too. The link between two people as friends is data that necessarily involves two people. Giving one of those people power to move that data gives them power to move data about someone else, too. Interoperability also creates security holes, making you only as secure as the least secure social network site. There are social nuances among users that shape the privacy and security of social network sites.
conclusions
The same social factors that make us use social network sites are the same that lead to making privacy mistakes, and are the same factors that lead to harms.
Privacy violations are peer-produced.
Facebook et al acts like a privacy violence, turning your social replications mechanisms against you and causing you to infect your friends, too.
questions and conversation
Q: Do we really want to save Facebook?
A: It’s not just about Facebook, it’s about human tendency to have social interactions and to form communities, and the way this unalterable tendency interacts with technology.
Q: What if the most privacy-conscious Facebook users were able to draft the privacy policy?
A: Facebook loves the publicity of turning this kind of thing over to their users, but it’s not likely that they’d actually give over that kind of power.
Even if Facebook is perfectly transparent about their operations, allowing us to keep a large corporation from beating up on a person, this doesn’t fix the problem of Facebook creating a platform that allows people to beat up on each other.
Q: Is it really plausible that people doing this don’t understand the nature of what you’re doing? If a user reads other people’s Walls, won’t they realize that their own posts are viewable to others?
A: If it’s dumb to make these kinds of decisions, then we should ask, “What makes people make such dumb decisions?” The match between what people expect and what they get isn’t very good, but there may be interface design elements that will allow us each to correct for that. Possibly: allow people to look at their own profiles through someone else’s virtual eyes. Take on the guise of a Total Stranger, or Co-Worker, and look at your own profile to set what will be viewable. This may be much more helpful in matching expectations with results than asking people to check a list of boxes to make those same settings.
Q: Look at it from a different perspective: it’s not that people shouldn’t post photos of them going out, it’s that employers shouldn’t be looking for this kind of thing. Potential for general privacy norms to shift over time, so that employers eventually won’t fire you for having photos of yourself doing keg stands?
A: It’s a dynamic effect. Even if we hit an equilibrium about privacy and reputation norms, these platforms are open to people who are too young to have internalized these norms, to make good decisions about this. There’s also a generational effect, where a small site becomes popular because it’s safe or private, and this eventually attracts enough users that these initial privacy expectations are no longer accurate.
Q: Blossoming of a new honor and shame culture – has anyone looked at this anthropologically? Reputation economy studies of Facebook would be really interesting.
A: Burger King had a promotion that they’d give you a free Whopper if you defriended 10 friends.
Q: Employers can look at Facebook profiles to find information that they’re not legally permitted to take into account when making hiring decisions. (Examples: potential for starting a family, minority status) How should we take this into account?
A: We need to think about spaces, and the relationship between virtual spaces and status in society. Social networks start out as a fairly privileged place (Friendster for upper middle class, Facebook for Ivy League college students). Who is the space for, are outsiders a threat? Will this create chilling effects on people’s expressive activity? How can you protect this? Interesting class divisions emerge among various social network platforms – MySpace is more high-schoolish and less buttoned-down, LinkedIn is for professionals, and Facebook is for preppy college kids/professionals-in-training.
Q: If you’re putting images of your babies on your profile, what result for your kids, and when they join Facebook?
A: Yeah, and all those bloggers who talk about their kids — will the kids demand that your blog archive be taken down when they come of age? Changes your relationship with your children.
Q: If we’re going for a more cautious set of web users, don’t we need the good examples of privacy crash-and-burn? Nothing better to convince me not to Ghost Ride than seeing that idiot run his truck into a tree!
A: Need high-profile bad mistakes and also good advice for how to avoid that. Want to minimize the number of high-profile crashes, to minimize the harm to people.
Q: What about a central index of Ways Facebook Has Screwed Up My Life?
A: Good idea to have an external ombudsman monitoring this kind of thing, and also have it as a resource for new users, to introduce them to the good and potential bad of the service.
Comments
Flickr Photos
|
A2K4 Update
yaleisp.org search
Recent Posts
- Governmental Deliberations on Exemptions to the DMCA’s Ban on Circumvention
- MFIA Wins Appeal Seeking Access to Sealed Records
- Innovate/Activate Unconference on September 24-25
- Bilski & the Definition of Things That “Are Free for All to Use”
- Illuminating the impact of intellectual property law on innovation