[sanswork]

I feel like they aren't even trying. The number of times I've reported obvious stolen accounts running scams or spamming only to recieve the "we investigated and found no rules broken" has made me stop trying. Every concert listing is full of scam bots posting the exact same wording to scam people.

Given the ability to shadowban from public posting and a few hours I'm pretty sure I could write a single function to block 95% of the scams. It would be one thing if they were dealing with complex scammers but the fact is they haven't even tried to stop the very low hanging fruit that you could solve with a few regexes.

[650REDHAIR]

I was perma-banned on Reddit for “abusing” the report feature.

Reporting illegal firearms sales.

The post and report in question were ~6mo old at the time of review and I believe had already been deleted by either the OP or a sub mod.

I got a similar warning (but no ban) reporting similar sales on FB.

They don’t care to fix it

[soco]

Maybe I should stop reporting all those charity scams on Tumblr, before they think I'm being their spammer...

[laweijfmvo]

haven’t they claimed that their “AI” catches 99% of spam before it gets seen or something? seems like there’s even lower hanging fruit or they’re just lying.

[mrguyorama]

I would bet money Facebook does very little fraudulent advertising filtering, since they are explicitly incentivized not to.

But "We block 99% of fraud/spam" is an outright useless and purposefully non-informative datapoint. It doesn't tell you anything. As a statement, it is equally true both for the SuperGoodMerchant who lets through a single fraudulent advertisement a year, and it is also true for UltraScamAds who let through a million scam ads a day. It depends entirely on the rate of fraud attempts compared to the rate of legitimate requests.

An actual useful metric if you want to avoid fraud or scams would be, what percentage of ads do they run that turn out to be fraudulent?

My anecdotal experience is like 95%.