Thank goodness I stopped using WhatsApp when I did – things are about to get MESSY. As of this Facebook currently has a content moderation plan*… another NEW content moderation plan I should say… Which will use AI to flag messages before they’re encrypted, sending said messages then to Facebook HQ where they can be read by humans. Messages will then be used by Facebook to make their moderator machines smarter and better at detecting term-violating words and phrases from their always-editable lists.
“New” content moderation plan
*Facebook hasn’t done particularly well in the public eye when it comes to their largely farcical content moderation system in the last few years. They’ve got another “new” plan in place this year because of this little slap on the wrist from the FTC.
So-called end-to-end encryption
As Bruce Schneier expanded upon, Facebook’s latest plan includes a bit of a paradox. WhatsApp is supposed to have end-to-end encryption, which SHOULD keep communications private and secure as such. But Facebook wants to add embedded content moderation and blacklist filtering algorithms to all their services, including WhatsApp.
Facebook’s vision for a better moderated platform will require an element mentioned in their recent public talk “Applying AI to Keep the Platform Safe.” This talk took place on May 1, 2019, at the 2019 edition of their developer conference F8. Facebook said in that speech that they will need to be able to scan contents of users’ messages before they’ve become encrypted.
Facebook also mentioned that they’ll need to harvest messages that’ve been marked by their AI as potentially breaking their speech rules. When a violation is flagged, said Facebook, they’ll need to be able to access those messages for human reading.
Once violation is found, the flagged messages will then almost certainly be used in their entirety to further train their algorithms. All without any sort of additional authorization on the part of the message sender or receiver.
Facebook will be able to list any set of words or phrases that’ll be used by its AI machines to seek out and flag. These machines appear as artificial intelligence algorithms in the WhatsApp app, locally, on your phone.
The WhatsApp app will read all messages as they’re typed on each person’s phone. If a violation is found, violator messages will be sent directly to Facebook, in their entirety, for further action.
Violating your trust
Potential for further forcing of action on the part of government systems AND malicious potential on the part of 3rd-party actors be damned, this system’s most BASIC elements are a level of dishonest and inappropriate that I’m ALMOST shocked Facebook found any bit of it acceptable for real public implementation.
WhatsApp’s transmitted messages may remain end-to-end encrypted, but that does NOT mean that everything you type or prepare to post with WhatsApp will remain private. That is simply no longer true, if it ever was true to begin with.
BONUS: To be clear, WhatsApp IS Facebook
At the announcement of acquisition for WhatsApp by Facebook, it was clear what Facebook was after. Zuckerberg literally said the following: “WhatsApp is on a path to connect 1 billion people. The services that reach that milestone are all incredibly valuable.” He went on to say that he was excited to “partner” with the service and that they would “make the world more open and connected.”
In early March of 2014, privacy groups gave a list of reasons why Facebook shouldn’t be allowed to acquire WhatsApp. While initial outcry seemed to fall by the wayside – at least in major news blasts – in the few years after that acquisition, things have begun to light up once again here in 2019.
A report in the Wall Street Journal tipped that the Federal Trade Commission (USA’s FTC) is currently examining Facebook’s acquisitions over the past 15 years as part of an antitrust investigation. They’ve suggested that the investigation is looking into whether the acquisitions were made “to snap up potential rivals to head off competitive threats.”
NOTE: WhatsApp is the subject of this article because WhatsApp was supposed to be end-to-end encrypted and private. That’s how it’s been promoted for quite a while now. If WhatsApp is subject to scanning, so is every other communication system and app in Facebook’s stables. That means Instagram and Facebook Messenger too!
UPDATE: Since this article was first published, several very important WhatsApp updates appeared, both in-app and in the outer realm of privacy – the Facebook realm. You’ll want to take a peek at how WhatsApp monetization is here – even if you don’t see it right up front and center. You’ll also want to see the latest on Facebook WhatsApp encryption plans re: US Government scrutiny – privacy-related things haven’t gotten better since this article was first released. If anything, they’ve gotten worse.