A Simple Solution to Improve Broken Peer Review System at AI Conferences [R]

“`html

• The proposed solution from a British AI enthusiast suggests dividing authors into two halves (A and B), ensuring that no author is both an author and a reviewer in their own half. This prevents the incentive for reviewers to reject good papers to increase their own chances of acceptance.
• Each author/paper would be placed in one of these halves randomly, with all authors belonging to the same half. The conference organizers can then assign reviewers from the other half to each group independently. This separation ensures that reviewers do not have a conflict of interest and are motivated by fairness rather than self-interest.

“`