I was briefly in contact with folks in a team at Amazon who said they were working on fixing the fake review problem. I ultimately didn't pursue a role there because I just don't really trust Amazon as an employer. (I worked there previously and got burned badly.)
Incidentally, I had also been in contact with Facebook about a role "combatting abuse" across their services. That loop was a bait and switch for standard ML (eg, improve video recommendations for users to keep their eyeballs on FB).
I try to assume good intentions, but I'm constantly given evidence that there are no good intentions in these companies. I don't think Amazon really has any reason to combat fake reviews. And I know a good number of engineers and scientists there. It's a tractable problem.