Agreed. I recently heard in a talk an example of NLP being used to help rank sellers by analyzing messages between buyers and sellers, before and after transactions.
Another example was how Taobao increased the number of useful reviews by allowing sellers to pay for reviews, and used NLP to determine if the review was useful. If it was, only then would the reviewer get paid. This would then dis-incentivize sellers expecting bad feedback from paying for reviews, and incentivize good sellers who were seeking honest, relevant feedback.
Seems like both examples can still be gamed, but it seems like a step in the right direction.
Another example was how Taobao increased the number of useful reviews by allowing sellers to pay for reviews, and used NLP to determine if the review was useful. If it was, only then would the reviewer get paid. This would then dis-incentivize sellers expecting bad feedback from paying for reviews, and incentivize good sellers who were seeking honest, relevant feedback.
Seems like both examples can still be gamed, but it seems like a step in the right direction.