Frances Haugen asks Government Pension Fund of Norway to sell out of Meta

> > The Nobel Peace Center advertised the event precisely on social media, but the ad on Facebook was removed, without further explanation, as "political advertising". The center has been able to advertise other events on Facebook without any problems.

> more inclined to just think that once again the ML enforcement models that the company uses have thrown a false positive

Sorry, but when Aunt Agatha shows up dead in the library, the butler who benefits from the will, and gives her medication, is first on the list.

That's why I think your analysis is itself problematic and part of what's going wrong here.

What do you mean by false positive? On what criteria, that you personally know about how Facebook selects what to censor, do you "just think" such apologetics are deserved?

I don't wish to pick on you alone, because it's a trope here that we rush to defend self-evidently rotten technologies and their application by crying "just a bug!", "just a glitch!". It's because we love the technology and mistake an attack on those who abuse it for an attack on what we identify with.

> more than there being some shady desire to silence the event

That's far too charitable. They have every motive to silence the event, and the means, and a kind of ever-fresh alibi that "the machine did it".

That must be one hell of a an ML model to single out an event where the peace center (usually not a red flag entity) is talking about Meta. I mean, that's extremely convenient. Such cases, a computer problem. Well, that's just force of nature, nothing we can do about it. Please move on. /s

I hate Facebook more than most, but I wouldn't automatically presume that the claims in the post are the result of malicious intent from human actors.

Much to everyone's surprise, over the past decade Facebook had real humans looking into reports whenever a basic algorithm couldn't make a simple determination. That has completely changed in the past two years and they've gone full ML -- and the results reflect that, terribly. Facebook is once again flooded with fake accounts pushing crypto scams and ads that are basically direct links to malware, and reporting those scams does absolutely nothing, because the scammers have always been (and likely always will be) smarter than any algorithm. Things are much like they were in the early days of Facebook, before human intervention cleaned most of that up.

The bigger concern I have with going fully automated, is the lack of security. I use Facebook Messenger to send long lists of keywords to a friend, so that I'll start (immediately) seeing ads for all of the startups related to those keywords. It's a great research tool, albeit incredibly disturbing and something that should probably be illegal. But there's a terrible side effect to this, which becomes very apparent when you get targeted by their algorithm for a particular keyword that gets abused by scammers.

In 2020 alone, I spent over 150 hours sending abuse reports to every hosting provider and domain registrar imaginable, all for the same guy using Facebook's ad system to distribute extortion malware under the guise of being a Windows desktop version of Clubhouse. Most of the companies responded within 72 hours, which I'd call somewhat acceptable, though I still wonder how many people lost their data or got conned out of money in each of those time periods. Facebook, however, never responded once or took action on any reports. It would take them weeks (sometimes months) to remove this person's ads, and would rarely ban the actual accounts. It finally "ended" when I repeatedly contacted their legal department, who also never responded (and why would they), but did seem to light a fire under someone on the security team.

And then he was back about four months later doing the exact same scam with different names, because all they did was add some filter for Clubhouse ads. He's still doing it, but it's an impossible battle if Facebook refuses to participate in the real world, so I've given up on reporting anything. It shouldn't have been my job to fix it in the first place.

The way such links get 'replaced' is by finding an English language source or waiting for one to materialize.

> Who has the right to throw a stone, here?

- Those who build medical technology that saves lives.

- Those who make educational software that helps raise our young.

- Those who work on the digital support for real world infrastructure that delivers our food, supplies our electricity, and runs our transport.

- Those who work on software for publishing, arts and entertainments which underpin our entire modern culture.

Need I go on? Please don't conflate the ethics of a whole industry with the actions of a relatively few grubby little guttersnipes who make money from near monopolies exploiting other peoples vanity and isolation.