Hacker News Reader

Comment by shshhdhs | Hacker News Reader

My wife just experienced something that makes me assume Facebook's bot/abuse algorithm is malfunctioning.

She just tried to "sell" something for free on Facebook Marketplace. Due to the popularity, she received 100+ messages in a minute -- the notifications were rolling on her phone.

Shortly thereafter, Facebook locked her out of her account for 72 hours and asked her to verify her identity. I guess they thought she was a bot? Which is strange because why would a bot want to receive 100 messages rather than sending them?

View on HN

Why do you think she was flagged as a bot? I think it's more likely she was flagged as a scammer. To receive such a high volume of inquiries suggests something unusual going on. In her case it sounds as innocent as giving away an item that is in high demand, but in many cases a similar pattern might be a deal that's "too good to be true."

To phish people. Bots (like on Tinder, dating websites, selling websites, etc) will make the same post to a ton of different groups, and then send the people who message them to phishing links.

Ugh, lockout must be pretty frustrating. Receiving one message could be worth as much as sending 1000 messages when valued by conversion to the spam end-goal. It's also a less obvious technique and one that's more difficult to block. Clearly the algorithm is not as precise as it should be, but I'm sure this mechanism is preventing a lot of abuse even though it blocks some normal use.

Ok, so I got locked out of my account last night for 72 hours. I have no idea why. I had to upload an image of myself for their team to verify. I wasn't doing anything odd whatsoever. Still waiting for my account to be reactivated.

Maybe they just wanted a picture of you?

Hey, it's less awkward than asking him what he's wearing

One can dream!

You're beautiful.

In that situation I'd probably giving my best "grumpy old man from Up is not amused" face for the camera to quickly get across that 'yes, it's me, now let me back in' or such.

You could use a Facebook bot as CC for a botnet. Such bot would initially receive a lot of messages without necessary send out their own, and also being something Facebook would want to ban.

A bit late to this party, but ALL machine learning algorithms are wrong some fraction of the time. It's not malfunctioning, it's a false positive.

What was the product?

Reply on news.ycombinator.com