The ProPublica article that the ones you saw are based on was flawed, and has been updated. https://twitter.com/propublica/status/1436054877663375372
>An "end-to-end" encrypted messaging platform could choose to, for example, perform automated AI-based content scanning of all messages on a device, then forward automatically flagged messages to the platform's cloud for further action. Ultimately, privacy-focused users must rely on policies and platform trust as heavily as they do on technological bullet points.
Which doesn't break E2EE technically, but it certainly breaks it in spirit. And yes, I understand that really any application could feasibly implement something like this, it's not in many peoples threat models, etc. However, if I had to bet on which company would implement such a feature, it would be FB.
It just felt sort of funny, seeing this only a few days after all of those articles were written. Of course there is no way FB weaved the whole system and documentation together in two days, but I can't help but roll my eyes slightly at the timing of their release.