Category Archives: Privacy

Netflix, Facebook, and Social Sharing

Posted on by

Just as Neil Richards’s The Perils of Social Reading (101 Georgetown Law Journal 689 (2013)) is out in final form, Netflix released its new social sharing features in partnership with that privacy protector, Facebook. Not that working with Google, Apple, or Microsoft would be much better. There may be things I am missing. But I don’t see how turning on this feature is wise given that it seems to require you to remember not to share in ways that make sharing a bit leakier than you may want.

Apparently one has to connect your Netflix account to Facebook to get the feature to work. The way it works after that link is made poses problems.

According to SlashGear two rows appear. One is called Friends’ Favorites tells you just that. Now, consider that the algorithm works in part by you rating movies. So if you want to signal that odd documentaries, disturbing art movies, guilty pleasures (this one may range from The Hangover to Twilight), are of interest, you should rate them highly. If you turn this on, are all old ratings shared? And cool! Now everyone knows that you think March of the Penguins and Die Hard are 5 stars. The other button:

is called “Watched By Your Friends,” and it consists of movies and shows that your friends have recently watched. It provides a list of all your Facebook friends who are on Netflix, and you can cycle through individual friends to see what they recently watched. This is an unfiltered list, meaning that it shows all the movies and TV shows that your friends have agreed to share.

Of course, you can control what you share and what you don’t want to share, so if there’s a movie or TV show that you watch, but you don’t want to share it with your friends, you can simply click on the “Don’t Share This” button under each item. Netflix is rolling out the feature over the next couple of days, and the company says that all US members will have access to Netflix social by the end of the week.

Right. So imagine you forget that your viewing habits are broadcast. And what about Roku or other streaming devices? How does one ensure that the “Don’t Share” button is used before the word goes out that you watched one, two, or three movies on drugs, sex, gay culture, how great guns are, etc.?

As Richards puts it, “the ways in which we set up the defaults for sharing matter a great deal. Our reader records implicate
our intellectual privacy—the protection of reading from surveillance and interference so that we can read freely, widely, and without inhibition.” So too for video and really any information consumption.

Facebook Links Online with Offline

Posted on by

The Atlantic runs a story (citing the Financial Times) about how Facebook is working with Datalogix to link online advertising on Facebook with offline purchases by consumers:

Advertisers have complained that Facebook doesn’t give them any way to see if ads lead to buying. This new partnership is their response, as it connects real-life buying with ads seen on the site. Specifically, the service links up the 70 million households worth of purchasing information that Datalogix has with these buyers’ Facebook profiles. Using that, they can compare the ads you see with the stuff you buy and tell advertisers whether their ads are working.

That is, using your Facebook E-mail address and connecting that with your store loyalty card e-mail address, Datalogix will be able to say to (for example, CVS): “Yes, X saw an ad for CVS on Facebook and then showed up at your store and bought it.” The Atlantic article hints at the ramifications, especially given the types of stores that we can guess are participants in the program, and also notes that opt-out is a particularly insidious way to begin to tie our online lives to our offline lives.

The Atlantic helpfully provides a link to the Datalogix site to opt-out of the “service” (links included below) and notes that even finding the link on Facebook is like a treasure hunt. While on that site, I noticed that Datalogix also provides a link to opt-out of all Datalogix related tracking. The main “informational” Datalogix link (with the sublinks to opt-out) is here. To log out of the new Facebook program, use this link (it is a cookie based system, which means if you regularly clear your cookies, you may want to “protect” these opt-out cookies so that they’re maintained over time). But instead of using that direct link, try going to this page to see who else is part of Datalogix tracking, and decide who if anyone you really want to allow to continue to do that (they also offer a form-based link to opt-out of all Datalogix advertising, though that link requires you to enter your real name and home address).

The Atlantic story ends with a gentle attack on opt-out schemes, emphasizing that Facebook should at least make the opt-out option easy to find (something it has not done).

There is nothing surprising about this development. Facebook has been at the forefront of trying to eek more information out of its users and then seeking to commodify that information in as quiet a way as possible. In the past these kinds of changes have received significant pushback from users. While user outrage may also arise here, it may also be useful for users to go to the Datalogix site and opt-out of all Datalogix targeted advertising (not just that related to Facebook). If advertising service providers see that joining up with Facebook in one of Facebook’s now infamous submarine “opt-out” changes results not only in bad press about that particular pairing, but also brings them to the attention of users who then opt-out of all of that provider’s services, maybe the providers themselves will drive Facebook to use opt-in rather than opt-out as the presumptive choice when partnerships develop in the future.

Maybe.

Social Media and Chat Monitoring

Posted on by

Suppose a system could help alert people to online sexual predators? Many might like that. But suppose that same system could allow people to look for gun purchasers, government critics, activists of any sort; what would we say then? The tension between these possibilities is before us. Mashable reports that Facebook and other platforms are now monitoring chats to see whether criminal activity is suspected. The article focuses on the child predator use case. Words are scanned for danger signals. Then “The software pays more attention to chats between users who don’t already have a well-established connection on the site and whose profile data indicate something may be wrong, such as a wide age gap. The scanning program is also ‘smart’ — it’s taught to keep an eye out for certain phrases found in the previously obtained chat records from criminals including sexual predators.” After a flag is raised a person decides whether to notify police. The other uses of such a system are not discussed in the article. Yet again, we smash our heads against the speech, security, privacy walls. I expect some protests and some support for the move. Blood may spill on old battlegrounds. Nonetheless, I think that the problems the practice creates merit the fight. The privacy harms and the speech harms mean that even if there are small “false positives” in the sexual predator realm, why a company gets to decide to notify police, how the system might be co-opted for other uses, and the affect on people’s ability to talk online should be sorted as social platforms start to implement monitoring systems.

Cool but I the privacy implications are unfortunate

Posted on by

Ever heard of Book Depository? It is book store. So what? So let’s dance! Oh no that was Caddyshack. So they have map of what books are being bought from them and where. It is mildly mesmerizing. It seems not such a big deal, but as I was watching a book was purchased in Saskatchewan and someone bought the infamous 50 Shades trilogy elsewhere. They don’t seem to leave the history of the map up. Still, I think I’d be less than thrilled that my purchase was surfaced with location.

Lend me your ears, no really. I need them to ID you.

Posted on by

Researcher Mark Nixon at the University of Southampton “believes that using photos of individual ears matched against a comparative database could be as distinctive a form of identification as fingerprints.”

According to the University’s news site the claim is that: “Using ears for identification has clear advantages over other kinds of biometric identification, as, once developed, the ear changes little throughout a person’s life. This provides a cradle-to-grave method of identification.”

Ok so they are not taking ears. The method involves cameras, scans, and techniques you may know about from facial recognition. This article has a little more detail. As an A.I. system it probably is pretty cool. Still, it sounds so odd that I wonder whether this work has considered the whole piercing, large gauge trend. I can imagine security that now requires removing ear decorations regardless of what they are made of. Also if really used for less invasive ID, will wearing earmuffs be cause to think someone is hiding or should we remember that folks get cold. For the sci-fi inclined, bet that a movie will entail cutting off an ear for identification just like past films have involved cutting off fingers and hands to fake an identity.