With all the deep learning algorithms probably running on FB's servers, how come they cannot automatically detect these crude attempts at scamming their users? It seems they all follow certain patterns - keywords ("virus", "critical system error", etc.), a general alarmist tone, and a call for action below.
I'm just guessing but their's probably some cat and mouse involved. If facebook starts filtering one social engineering vector then they'll just shift to another one. It's like banning bots in various video games, you do it in random intervals in waves that obfuscate the exact cause and align nicely with the company's bottom line and PR efforts.
