Your messages are encrypted on WhatsApp… unless they are “problematic”

A long investigation of ProPublica details the work of WhatsApp moderators, staff outside Facebook who have access to data presented as end-to-end encrypted by the messaging service.

Since its switch to end-to-end encryption in 2016, WhatsApp has done everything in its power to promote its inviolable and ultra-secure side to users. In a few months, this house of cards has broken down somewhat. After the imbroglio at the start of the year around the gateway between WhatsApp data and Facebook, its parent company, a major investigation by the US media ProPublica comes again to cast shame on the messaging app.

According to this long article, message encryption would be bypassed in certain very specific cases. When you interact with your interlocutors, they have the opportunity to report a problematic message in their eyes. When this process is initiated, an unencrypted copy of the message is sent to a moderator, along with your five most recent messages.

Providers to read your messages

The catch is that these moderators don’t work directly at WhatsApp internally, being contract workers based in Austin, Dublin and Singapore. There would be around 1000 of these providers who would review the millions of content flagged each month and sorted by the messaging algorithm. Content that is supposed to be private, which may therefore no longer be so in certain circumstances. Obviously, the scope of this investigation must be balanced with the sad need for moderation on this kind of application. It is supposed to protect its users from all the excesses that hang out on the Internet such as harassment, death threats, child pornography …

According to the survey, WhatsApp moderators have profiles very similar to those of Facebook. Men and women aged 20 to 30, doing odd jobs and hired by Accenture, the company designated as an intermediary by Mark Zuckerberg’s group. “Sitting in front of computers in blocks organized by assignments, these workers use special software from Facebook to sift through the streams of private messages, images and videos that have been flagged by users as inappropriate. These entrepreneurs pass judgment on anything that appears on their screens, whether it is fraud, spam, child pornography or a potential terrorist plot, usually in less than a minute ”, can we read in the long article of ProPublica.