Algorithmic Pricing, Recommendation Systems, and Competition

AI-powered pricing algorithms raise concerns about supracompetitive outcomes without explicit coordination. Meanwhile, digital platforms use recommendation systems (RSs) to influence product visibility. This paper models Bertrand-Markov price competition in a differentiated product market with heterogeneous consumers, where both sellers’ pricing and the platform’s recommendations are AI-driven. The findings show that RSs can autonomously inhibit algorithmic anticompetitive conduct, resulting in prices even below the Bertrand-Nash benchmark. The results hold when the platform only prioritizes profits, as well as with variations in consumer heterogeneity, market conditions, and underlying learning parameters.