Recommender systems have innate bias, because they can only recommend items inside their scope of data; that and a reliable recommender system such as amazon, unless given specific instruction otherwise, will choose otherwise identical items and ratings with more ratings and/or clicks. Making it more difficult for already existing competitors to join the market late; aside from these innate biases in the RecSys; the RecSys suffers from bias in their data.
This article talks about how in an attempt to get new products into the market (combatting what I stated above) they would give products free of charge to customers, in return for their review. [However ReviewMeta found that these reviewers were “12 times less likely to give a 1-star rating than non-incentivized reviews, and almost 4 times less likely to leave a critical review in general.] This is an obvious bias that corrupts otherwise decent/good data; and corrupting any view of an otherwise great product recommendation algorithim.