BETA
This is a BETA experience. You may opt-out by clicking here

More From Forbes

Edit Story

Would Facebook's Move To The Edge Really Solve The Privacy Problem?

Following
This article is more than 4 years old.

Getty

As Facebook invests heavily in moving its content moderation to the edge, running its hashing and AI filtering algorithms directly on users’ phones, this approach could actually solve many of the company’s most pressing privacy issues, especially those involving data harvesters and intrusive governments. Once a user’s interest and behavioral data is stored in encrypted form directly on their device and out of Facebook’s reach, it will be much like Apple in its inability to respond to government data requests. Could this be the ultimate solution to its privacy problem or will harvesters and governments merely adapt?

Facebook’s move to the edge brings with it many benefits. Most strikingly from a privacy perspective, Facebook itself will no longer have access to the contents of a growing percentage of its user’s communications, as those are shielded behind end-to-end encryption.

Data harvesters will have little to show for their efforts when users increasingly communicate in small groups rather than publicly broadcasting their thoughts to the world like Twitter.

Most importantly, governments will no longer be able to turn to Facebook to request exports of a given user’s private information or to request a master database of the identities of everyone in their country that matches certain advertising selectors, such as being LGBT.

Much like Apple cannot penetrate the encrypted local contents of its users’ devices, Facebook will no longer have to worry about complying with government data requests, since all of the requested data will now be local on users’ phones, rather than sitting in Facebook’s data servers.

In fact, with the move to the edge, Facebook could even contemplate encrypting all small group posts, requiring every new group member to be vetted by each other group member to gain access to decrypt their respective content. This would prevent government officials and harvesters from accessing even group contents with the permission of each poster to see their respective content.

A repressive regime that submits a lawful court order to Facebook today requesting the identities and IP addresses of every person in their country that is LGBT and where being LGBT carries the death sentence, can readily receive such a list from the company. This is possible because all of the behavioral and interest data necessary to power its ad machine is stored in Facebook’s centralized datacenters.

In an edge-based future, those models would reside exclusively on each user’s device. As the user consumes and engages with content and posts their own content, their on-device ad engine would watch their every action, building up a model of their interests and using that model to make ad requests back to Facebook’s servers to display relevant ads.

Since this interest model never leaves the user’s phone, Facebook would have no access to it. Governments would be forced to go to each user and compel them to voluntarily unlock their phone or find other means of cracking the local phone’s authentication, much as they must do today with Apple devices.

Facebook would largely be rid of its need to service government requests.

Or would it?

In reality, Facebook’s datacenters would still be recording a log for each user of all of the interest-based advertisements they saw. Simply by looking at the advertising selectors on these ads, it would be trivial to determine which users have which interests, just like the company can do today. While the resolution might not be quite as precise, governments will still be able to perform most of the same interest-based surveillance using this advertising residue.

In fact, one key benefit for the company may be that under certain data protection laws, advertising log files may not be considered sensitive user data and thus exempt the company from many of its obligations under certain current laws.

Similarly, data harvesters will simply move to phone-based harvesting, building malware apps that try to intercept user input and display from Facebook’s app or offering to “manage” their posting behavior or give them a better viewing “experience” in order to intercept their Facebook activity. Though, Facebook and phone manufacturers could more easily combat such applications than they can today.

The company did not respond to a request for comment regarding its edge initiative.

Putting this all together, at first glance, Facebook’s move to the edge would seem to solve many of its most pressing privacy problems, dramatically curtailing data harvesters and government data requests. While data harvesters would likely find their activities partially curtailed, governments would still be able to leverage Facebook’s centralized advertising logs to determine which users have which interests, much as they do today. Worse, by creating a streamlined pipeline to deploy content filters to the edge on demand, Facebook will enable repressive governments to censor content they dislike in realtime or even turn their citizens’ phones into a massive mobile surveillance network.

In the end, it seems Facebook’s move to the edge will create a useful public relations story for it but do little to constrain the ability of governments to repurpose its global surveillance empire to Orwellian ends.