We develop a theoretical model to study strategic interactions between adaptive learning algorithms. Applying continuous-time techniques, we uncover the mechanism responsible for collusion between Artificial Intelligence algorithms documented by recent experimental evidence. We show that spontaneous coupling between the algorithms' estimates leads to periodic coordination on actions that are more profitable than static Nash equilibria. We provide a sufficient condition under which this coupling is guaranteed to disappear, and algorithms learn to play undominated strategies. We apply our results to interpret and complement experimental findings in the literature, and to the design of learning-robust strategy-proof mechanisms. We show that ex-post feedback provision guarantees robustness to the presence of learning agents. We fully characterize the optimal learning-robust mechanisms: they are menu mechanisms.

Maritime transportation is essential for global supply chains, but the issue of ballasting—vessels traveling without cargo—imposes significant economic and environmental costs. This paper focuses on the oil transportation industry, where about half of the total traveled miles are sailed empty, and reveals that fragmentation is the most important cause to ballasting after demand imbalances, accounting for 17-20\% of the total. We find that it is possible to reduce carbon emissions associated with ballasting by as much as 13\% by consolidating the market into small shipping pools, which avoids concerns about excessive market power. Consolidation improves utilization because larger pools better coordinate and diversify the set of ports they serve, which reduces the need for vessel relocations. At a higher level, this work shows the extent of the sustainability gains that can be obtained solely by organizing more efficiently the resources available in today’s supply chains.