While Europe has just passed the Digital Services Act, engineers at Meta (Facebook) warn of a very worrying point: the company simply won’t be able to say in 2022 where all user data comes from and where it is stored (physical data center ), making it difficult or even impossible to determine which jurisdiction these data depend on. The data infrastructure of social networks and meta-applications is designed according to the so-called “open borders” principle.
Meta’s systems collect, store and consolidate user data from a variety of sources – which may or may not be under Meta’s control (when third parties are involved). Once the data is consolidated, there is currently no way to determine whether or not all of this data comes from Facebook, another social network, online sales sites, and other ad trackers. In a report, engineers at the company are concerned about the choice in a context that appears to go against the stream of infrastructure axiom previously espoused by the company’s Mark Zuckerberg.
Facebook is not technically able to protect your data as required by the European Union
It states: “We do not have an adequate level of control or ability to explain how our systems use data, and as a result we cannot implement policy changes in a controlled manner or respond to external commitments of the type ‘We will not use data’ X for a purpose.” Y’. And yet it is now exactly what regulators are asking of us, increasing the risk of error and misstatement,” the report explains.
The engineering team therefore proposes a new system called Purpose Policy Framework (PPF) that would allow adding essential information to user data to allow some traceability. However, we already point out that this system only deals with data managed directly by Facebook, without worrying about data generated / collected / used by third parties. The whole question is whether this system will be effective enough to allow Facebook to comply with the laws outside of the United States.
And also to know if the system in question will be technically easy to implement in a company that already has a gigantic base of several billion users in the world – and all the more data that needs to be labeled correctly.
