Their defense is that they only process two million fraudulent transactions a year?
For example, my partner posted an item for sale in FB marketplace.
The buyer said she couldnt pick up item, but her brother would. She could pay via Zelle, all she needed was an email & phone # . None of that seemed weird, so my partner agreed.
The buyer then said she sent a Zelle payment. My partner got an email from "Zelle Pay" saying money was pending due to my partners account not being a business account.
Luckily, we know zelle and didnt fall for it. The sender was a Gmail account.
My partner must have been talking with a bot.
5 min later, another buyer showed up. She also could not pick up but her brother would. She was to pay via zelle.
Be careful with FB marketplace. It's a mess now.
I'm not sure there's a particular reason to think it's a bot or AI is there? It could just be a script and a call-centre type operation.
In my experience, in a FB marketplace transaction where the buyer picks up the item, the buyer only pays after personally inspecting the item and (often) haggling a bit.
Prepaying using Zelle for an unseen item is very weird.
I did this a few years ago to sell a PS5 in the outskirts of Metro Detroit.
We were in communication with the possible buyers and saw them turn into the lot. As soon as we asked them to meet us inside they got cagey, hung up, and just left.
Now? I think you've missed that by a bit which is understated in the same manner your "now" is.
In contrast, in Craigslist-land I'm interacting with just a phone number or an email.
Anyways, I sold a ton of stuff on FB marketplace before we moved out of our old place, only allowed cash / venmo payments and they had to come pick it up
This is the same reason why a PSP will always give you better fraud protection than rolling your own, they look at so much more data than you ever will that they will be able to better tell friend from foe. Fewer false positives and fewer false negatives. Still too much, but the theory is that in order to be able to detect fraud you will have to have some fraud!
That fails to argue that having any fraud is beneficial; if you have no fraud, there is no value in being able to detect fraud.
That fraud by itself is not beneficial goes without saying.
Just like any classification system a fraud control system has false positives, false negatives, true positives and true negatives. In an ideal version of such a system you have no false positives and false negatives. But that idealized version of that system (the one that perfectly classifies fraudulent transactions and non-fraudulent transactions) is static. It can't adapt to cases of fraud that it hasn't seen yet. Eventually a fraudulent transaction gets through.
And that transaction if not detected timely but after the fact (so now there is some detected fraud, in the interval prior to its detection it was undetected fraud) will result in an update to the fraud detection system (probably it will need more than one example to be able to generalize from). Now the perfect system may start to generate false positives or false negatives on other transactions.
Run the 'perfect' system for long enough and it will be like every other fraud control system, a certain fraction of fraud will get through and a certain fraction of good transactions will be marked as fraud. How low you can do this sustainably is an open question but there is a lot of money to be made by squeezing out the last little bits and so a lot of money is poured into these systems. The nasty bit is that many of the players in this space attempt to off-load their residual risk on others (such as merchants or consumers) and that gives them less of an incentive to do the best job they can.
So that's why all of this is relevant. I hope I'm able to explain it clearly enough.
It's like the people who used to abuse the friends/family policy on paypal to cut out paypals fees and then were surprised when they had no buyer/seller protection, when they had to click through several messages specifically telling them that.