The New York Times has once again gotten its hands on a cache of documents from inside Facebook, this time detailing data-sharing arrangements between the company and other corporations, which had “more intrusive access to users’ personal data than [Facebook] has disclosed” for most of the past decade, the article revealed.
Microsoft’s search engine, Bing, got Facebook users’ friends, whether or not the users agreed to grant that access. Netflix and Spotify got access to users’ messages. Amazon got names and contact information. And, of course, Facebook got things in return. The Times states that Facebook used data from other companies, including Amazon, in its “People You May Know” feature, which has long attracted attention for its mysterious suggestions.
But while the story recalls the explosive Cambridge Analytica episode, it’s much more mundane. These were not bad actors, but merely actors playing exactly the role that Facebook wanted them to play. The goals of these integrations were not nefarious, at least from what we currently know, even if the idea that Spotify’s engineers would have access to your Facebook message data is probably not intuitive to most people.
Facebook responded to the story with a long blog post in which the company argued that the data-sharing “work was about helping people” do things on the internet “like seeing recommendations from their Facebook friends—on other popular apps and websites, like Netflix, The New York Times, Pandora, and Spotify.”
Which, sure: That was one thing that these data-sharing partnerships allowed. But they also allowed Facebook to grow, and grow, and grow. To entrench itself everywhere in the social-media ecosystem. Facebook was happy to trade user data to expand its business operations, and to pretend that this was all about users defies reality. Users got a small “improvement” that they didn’t ask for. Facebook got permits to build the pipes underlying its data empire.
Back when the data-sharing partnerships began, in 2010, the vision Facebook had of itself could be called Everything-but-With-Facebook. The service would be the social spine for all other services on the web. You’d log in with it, share through it, integrate your Facebook friends into all online experiences. This vision had an arc that began with integrating Facebook with also-ran phone makers and ended in the failure of the concept, overall. But in between, as The Verge’s Casey Newton points out, it gave away more and more data until it overreached with what it called “instant personalization,” which customized results in Bing with Facebook data.
The company has been pulling back on this kind of arrangement for years now. It admits in the Times story, however, that the change was not primarily because of privacy concerns. Most of the deals that Facebook cut simply didn’t work for either party, despite the data transport going back and forth. As Android and iOS took over from the wider world of mobile phones and computers, Facebook’s vision of what it should be evolved. It would no longer be the social spine, but the suite of apps you cannot escape. For years now, the model has been: everything inside Facebook. Apps that threatened that hegemony were purchased (WhatsApp, Instagram) or battled tooth and nail (Twitter, Snapchat).
What’s fascinating is that, as with Cambridge Analytica, we’re mostly talking about the sins of Facebook past, remnants of a different idea of how the internet was going to work. Except that the Times’ reporting indicates that data access for many companies continued long after it should nominally have been cut off. Other companies purported to be surprised that they had the depth of access that they had. The sloppiness—basically up to the present day—remains the most incomprehensible part. For a company that is user data, Facebook sure has made a lot of mistakes spreading it around.
By the looks of it, other tech players have been happy to let Facebook get beaten up while their practices went unexamined. And then, in this one story, the radioactivity of Facebook’s data hoard spread basically across the industry. There is a data-industrial complex, and this is what it looked like.