> But if it were so easy, why do companies not work harder to block them out?
Heuristics that are easy for a human to apply aren't always so easy for a computer. How do you quickly judge in a programming language if a photo is of professional quality but nothing about the renting agent is mentioned? How do you judge if it's priced high enough to be vaguely plausible, but suspicious in context, by knowing intimate details of the local market in the way someone living there does?
I think detecting scams is much harder than most people really want to admit. Which is why stopping them is challenging.
Is it the liability issues (blocking content implies taking responsibility for the content instead of just being a middleman)? Is it fear of driving the scammers to make ever better listings? Is it that the costs to do this outweigh gains? Is that the tech would have too may false positives?
All of these 4 steps are technically automatible; what's stopping these companies from implementing them?