Recent reports from the AAAI-27 review cycle have raised concerns about reviewers coordinating bids for reciprocal assignment advantage. Prior work treated bidding, reviewer assignment, and review manipulation as separate stages, leaving the lifecycle effects of collusive bidding unclear. To address this gap, the authors introduce CABAL, an end-to-end multi-agent simulacra framework that holds the conference environment fixed and configures LLM-driven reviewer agents with either honest or collusive policies.

The framework includes an affinity-guided collusive bidding strategy that uses mutual reviewer-paper affinities to construct collusion rings and select target papers, producing expertise-consistent rather than arbitrarily targeted attacks. Controlled experiments show that collusive bidding more than doubles target-paper capture, and assigned colluders score target papers about two points higher than honest co-reviewers. Conference-wide effects, however, remain comparatively modest.

Evaluated bid-phase detectors provide only limited evidence of collusion. In a fixed-triplet detector stress test, native positive-bid graphs are confounded by benign affinity, while a Very-High-only diagnostic view enables precise but low-coverage local recovery. The results suggest that detecting collusive bidding in real-world peer review remains a difficult challenge, even when the attack strategy is known to the evaluator.