We're looking at a lot of ways to detect simply "echoed" thoughts. For example, one method we're looking at is having sources be able to report one or more reasons for their opinion. This makes it possible to know better how to mathematically evaluate an opinion.

For example, if the "reason" reported is that the opinion probability comes from a new scientific experiment (new in the since that it hasn't previously been included as evidence in the probability calculation), then Bayes algorithm is a good algorithm (although even in this case the new evidence can be "watered down" by trust factors).