From the white sheet.

Peer Review
The contributor’s workflow begins with adding or editing information in the knowledge
base in hopes of receiving rewards. Each contributor’s submission costs gas, so
flooding the system with junk has its economic limits, just as in Ethereum.
​
All
submissions go through a mandatory peer review process and are not committed
to the knowledge base unless approved.
​
Every submission requires sources for
reviewers to validate the submission’s reliability and accuracy. Hence, spam, fake, or
malicious content will be filtered out.
Every user who submits a contribution is also required to peer review other
submissions. We use Machine Learning to match peer reviewers with topics with which
they will likely be familiar. Submissions are not committed unless they pass the
judgment of multiple peer reviewers.
It is crucial that peer reviewers are also contributors to ensure that only individuals who
are economically aligned with the well being of the ecosystem are able to peer review
submissions. Bad actors who try to attack the system will be fighting against legions of
good actors driven by a powerful incentive system.