If we got rid of all of the bots, that wouldn't make Amazon easier for me to use. I feel like scammers have already figured out that humans themselves are relatively cheap to buy.
Bots aside, what we post on the internet (and believe in general) is a function of the information we consume. That information is increasingly consumed through the internet, and the internet (as typically used) is feeding us increasingly poor quality information. (To be consumed, then shared. Recurse...)
We can (and I think do) cope by having different levels of trust for different sorts of internet information. I generally have higher trust in an HN post than an Amazon review, for example. This is useful, but has a dark side: some of the things I (and likely, you) take as a sign to "increase trust" happen to correlate with things that have nothing to do with trust. I trust Amazon reviews more if they have good grammar and spelling, largely because I trust general internet information more when it has good grammar and spelling. But perhaps I shouldn't when e.g. buying a screwdriver (what does writing ability -- and the things it correlates with -- have to do with evaluating a screwdriver?). Coming up with social/political-flavored-information examples is an interesting (and worrying) exercise.
Consider this: a consumer that employs above average grammar and spelling skills to write product reviews may also be more skillful in forming and expressing valuable opinions and assessments of any product.
I think that people are naturally inclined to pick up on those signals, whether the correlation is real or not.
back to the Amazon example: an expert car mechanic may have worse spelling/grammar than a hobbyist, but I should probably trust their Amazon review of a car part more.
[1] https://opinionator.blogs.nytimes.com/2012/08/08/hear-all-ye...