Calling them 'consumers' rather than 'people' seems like a use of dehumanizing language to assuage your guilt. At least, that's how it comes across to me. When you accept a bribe for a review you're screwing over the living breathing human beings who might make the mistake of assuming you aren't a liar.
Amazon does offer a legitimate program for brands to offer free gifts in return for reviews, but it doesn't require those reviews to be positive.
If you follow Kant, the argument is pretty much this: if everyone would do that, nobody could trust any reviews. The right choice, according to Kant, is that which you can want to be made into law for everyone. You don't want reviews to lose all meaning as they're important information for people to learn about the products, ergo you mustn't fake it.
Should you blame them? Yeah, but it's a pretty small infraction and they're a small cog in a large machine, so don't be too harsh.
The question should be whether or not the embodied process is fair, in whole, to the users? As much as we worry about "right" API design here (or what is the perfect framework for CSS?), you'd think we'd worry about how bad we are screwing users with all this software optimized to make startups, founders, VCs and private equity firms more an dmore money with less and less effort.