Everything from simple sentiment analysis, to archive.org, to another mirror. I hope that does not discourage you from releasing the data.

Edit: I see the other comment about archive team already collecting and releasing this data, for free in an open format. I think that will be a good first source as well.

I think the priority was to reenable discussion of new movies and TV shows, that would be useful moving forward. But maybe they could make an API.

I'd love to explore how this data could be used to enhance recommender systems.

AFAIK, Jinni is doing exactly that.