Chat: we do nothing on complex objects (notably clearly violating groups & pages)
Thread View
████ name redacted. Different bars may be the same person or different people.
quoted in an excerpt in the archive.
Thursday 1 October 2020
07:13Antigone Deborah DavisHey did you ask for more child safety specific HC? If you have 5 minutes for a call, I can provide some context for the question.
Christopher
07:22████yep i can drop out now Christopher
07:22████will call you Christopher
07:23████free for the next 7 mins
09:24Antigone Deborah DavisHere is the draft deck I have put together for Nick. would love your feedback. This is really a first conversation on his part.
09:24Antigone Deborah Davishttps://drive.google.com/file/d/17GNUEUgjEaIUUjclrRTuiqDgFa-b1U8s/view?usp=sharing Christopher
09:53████would you benefit from specific examples of where we fall short?
eg:
- we do nothing on complex objects (notably clearly violating groups & pages) - we can't detect 'new' CEI - we're scrabbling to be ready for encryption - our IIC numbers (notably on IG) are through the roof - we don't have comprehensive logging or measurement, we need both - anything beyond *sexual abuse is effectively being ignored - we don't have bandwidth to properly support new products (i worry about Reels & RTC in particular) - we have big surface gaps, reporting gaps, detection gaps, and review capacity gaps Christopher
09:54████(that's not me complaining, just some top-of-mind worries) Christopher
09:55████the slide on headcount vs health & privacy is really strong
09:56Antigone Deborah DavisSure, some of this in the original quip doc we put together https://fb.quip.com/QRndAZrYZuwc, would be great for you to add your two cents there. Or I can just pull from the work chat and add it and save you the time.
09:56Antigone Deborah DavisIt is quite astonishing.
09:58Antigone Deborah DavisHeard the CEI solicitation classifier made it through LAMA.
09:58Antigone Deborah DavisWahoo!
Christopher
10:00████we got 2 things through:
1) blocklist for 'obvious' terms (like 'kid porn' etc) 2) classifier for auto-enforcing at high precision on a wider pool of terms (most will still be fairly 'obvious') Christopher
10:00████(it's actually a different classifier to the CEIS one a specific keyword classifier
10:27Antigone Deborah DavisAhhh PLTF EX