Facebook has been caught archiving videos users thought they had deleted. The New York Magazine flagged the problem to the company last week, after a user downloaded their Facebook archive and was surprised to discover multiple takes of a video they had thought had been discarded at the time the recording was made, years earlier.
Facebook has now apologized for failing to delete the videos, saying “a bug” prevented draft videos from being deleted. It adds that it’s in the process of deleting the content now.
A company spokesperson told us: “We investigated a report that some people were seeing their old draft videos when they accessed their information from our Download Your Information tool. We discovered a bug that prevented draft videos from being deleted. We are deleting them and apologize for the inconvenience.”
It has also confirmed the deleted content was never publicly shared, although it’s less clear whether Facebook’s systems used the outtakes in any other way — such as another signal for its ad targeting systems, for example.
The social network’s business model is based on building detailed profiles of users so they can be segmented into custom audiences for advertisers to pay to reach.
The undeleted videos were recorded via an old feature that allowed Facebook users to make and post videos directly from their web browser. This system functioned by streaming the videos to Facebook as they were being recorded — hence, without a working delete function being in place, it was possible for reams of outtakes to be unintentionally archived on Facebook’s servers.
Having proper systems in place to delete personal data on request will be essential for Facebook’s compliance with the EU’s incoming General Data Protection Regulation (GDPR) — or the company risks massive fines.
While Facebook is ‘fessing up to having stored data that users believed they had deleted in this instance, it has previously denied that it collects unposted status updates — after rumors about Facebook keylogging anything users input into the platform swirled in 2013.
That was following the publication of a Slate article about a research paper written by a Facebook data scientist and an intern examining moments when users started to write a status but subsequently did not post it.
“Facebook does not collect or track any content that people have chosen not to post,” a Facebook spokesperson told Mashable in early 2014, in response to the status update logging rumors. (Albeit that statement is not accurate in the case of the archived draft videos — so it would probably have been more sensible if it had also included the word “intentionally”.)
In the 2013 study, Facebook users whose status writing activity was looked at for the research were anonymized (“researchers were not privy to any specific user’s activity”, the paper notes); and content of any self-censored posts/comments was not sent back to the company’s servers — rather only a binary value indicating that content was entered was sent.