Have you downloaded your Facebook data archive yet? Facebook makes it easy to obtain a ZIP file of all the data it has on you: your status updates, your friend list, your messages … and, as I and several people I spoke with were surprised to discover, every video you ever filmed on the platform — including videos you never published.
Cambridge Analytica’s effect on the election is still disputed, but if there’s one thing everyone can agree is true in the wake of the scandal over its data-harvesting, it’s that we’ve paid far too little attention to the scale and scope of the data that Facebook has collected on us over our years of platform use. It’s true that, in most cases, we’ve “consented” to this kind of data-gathering by clicking “okay” on a dense terms-of-service contract, and neglecting to opt out. But given how often the details are ignored, it’s disconcerting to discover exactly what data has been gathered.
Last week, Facebook users around the world discovered that Facebook’s Android app had, in some cases, Hoovered up extensive call data without their awareness. “When this feature is enabled, uploading your contacts also allows us to use information like when a call or text was made or received. This feature does not collect the content of your calls or text messages,” Facebook said, noting that users had to initially opt in to have their communication tracked. But it seems unlikely that many of the people who “opted in” were aware that they’d done so, or what exactly it meant.
And call logs aren’t the only data Facebook’s been holding to user surprise. Earlier this week, like many people around the world, my sister Bailey downloaded her Facebook data archives. Along with the contact lists and relationship statuses was something unexpected: several different videos of her attempting to play a scale on a wooden flute in her childhood bedroom.
Each video, she discovered, was a different “take” — recorded on Facebook, but then, she assumed, discarded before she posted the final version to a friend’s wall. (In the archive, you can infer which video was posted on a wall versus the ones that weren’t based on the comments. Videos that were never published are commentless.) In one of the clips, you can hear Kircher say, exasperatedly, that it is her 13th take. At the end of the clip, which isn’t to her liking, she groans and reaches forward, apparently to delete the video and try again.
How did this happen? In the pre–Facebook Live era, leaving videos on your friends’ walls was something of a crude FaceTime. You’d post a video, they’d respond with a video, and so forth. Importantly, Facebook had a feature that let users film videos via webcam on Facebook itself — that is, without ever leaving the Facebook site to use a video recorder. Once you were done filming, Facebook would show you a preview of your clip. If you decided to do another take, you could click to discard that video and try it again. Except, the video wasn’t actually deleted. Instead, Facebook apparently saved your unused clip.
Another co-worker, Brittany Stephanis, found over 100 videos in her archive and says that she only ever publicly posted about a third of them. The earliest date back to Christmas Day in 2008 when Stephanis, then 13, started recording videos to wish her friends a happy holiday. Stephanis says that her archive contains videos she clearly never planned to shared with anyone. “There are videos of me just checking my teeth,” she explained. My sister also had videos — rehearsing for school musicals and cheerleading — where she was using Facebook’s desktop camera to review herself and then erase, or so she thought, the video forever.
I thought this was a funny quirk of my sister’s. And then I looked at my own data archive. There, at the bottom of the list of videos I’d put on the platform, I found clips that appear to have never been posted to Facebook but were saved anyway.
From what I can tell, most people haven’t noticed this yet. (Or, possibly, they have and are unperturbed.) One reason might be that the videos are saved in the FLV format, short for Flash Video. Most default video players can’t read them, but you can watch them in VLC. You can’t preview a .flv file in a file browser, which could make it a little trickier to figure out what videos Facebook has of yours. To find yours, download your data archive — instructions here — and open the file. Select the index, which will open your archive in your web browser. From there, click the video section to see what’s what. (You might need to download the free app VLC player, which can play Flash videos.)
Facebook saving data you didn’t share isn’t an entirely new concern. Back in 2013, then–Facebook intern and Ph.D. student at Carnegie Mellon Sauvik Das and Adam Kramer, a Facebook data scientist, put together a study on what they deemed “self-censorship” on the platform. Self-censorship, according to their work, was anything a user typed in the status box but ultimately didn’t post.Data that Facebook can track. Facebook’s current data policy says that the company can “collect the content and other information you provide when you use our Services, including when you sign up for an account, create or share, and message or communicate with others.” “Create” is the operative word in there. By that logic, Facebook technically could save any video a user filmed but did not publish because you created it on the platform. Still, that requires the kind of close read of the fine print that most humans, at least not before Facebook’s ongoing Cambridge Analytica privacy fiasco, likely aren’t taking the time to do.
An industry expert recently told me about another “open secret” — that when Instagram first introduced video, the company would begin uploading a user’s video while they wrote a caption. That way, when the user pressed “post,” the video would already be good to go, and the user wouldn’t have to wait while the video uploaded in real time. Instagram co-founder Mike Krieger explained this in a presentation on the “Secrets to Lightning Fast Mobile Design” in 2011. It doesn’t seem unreasonable that a similar system might have been used at Facebook to create a faster, cleaner user experience. Except, unlike Instagram which says it threw out the data if a user cancelled the post, it appears Facebook was keeping the videos.
I reached out to Facebook for comment today, and the company asked me to connect them with Kircher so she could consent to the company investigating why these outtakes were in her data archives. The representative told me the company thinks they’ve “gotten to the bottom of it,” but that Facebook was still figuring out what it could share. Later, a Facebook representative provided a statement saying that the company is still looking into the matter: “We’ve heard that when accessing their information from our Download Your Information tool, some people are seeing their old videos that do not appear on their profile or Activity Log. We are investigating.”
The videos that Stephanis and Kircher found in their archives are pretty harmless, at least in terms of content. But it’s easy to imagine a user filming something that wasn’t quite so juvenile and being horrified to discover that Facebook has a copy. Apropos of, well, everything, now seems like a good time to mention that Mark Zuckerberg keeps a piece of tape over his webcam. Seems like he might be onto something.
Update, March 28, 6:16 p.m.: This piece has been edited to include a statement from Facebook.
If you’ve also found questionable things in your Facebook data archive, I’d love to hear about it. Please email me at firstname.lastname@example.org.