This is a research prototype. The data and analyses are preliminary and not yet validated — we'd welcome your .
Two third-party Amazon sellers employed automated pricing algorithms that lacked sanity checks, resulting in a recursive price-hike loop. The algorithms continuously adjusted prices in response to one another, inflating the cost of a biology reference book to over $23 million before the sellers manually intervened.
Amazon's pricing algorithm was implicated in a reference book about flies' unusual high price of millions of dollars, allegedly due to two sellers using the paid service which based their product's pricing on one another's as competitors.
Domain classification, causal taxonomy, severity scores, and national security assessments were LLM-classified and may contain errors.
AI systems that fail to perform reliably or effectively under varying conditions, exposing them to errors and failures that can have significant consequences, especially in critical applications or areas that require moral reasoning.
The incident was caused by a lack of robustness in the pricing algorithms, specifically the absence of sanity checks or price ceilings to prevent an infinite feedback loop.
AI system
Due to a decision or action made by an AI system
Unintentional
Due to an unexpected outcome from pursuing a goal
Post-deployment
Occurring after the AI model has been trained and deployed
The pricing anomaly was caused by the autonomous interaction of two deployed pricing algorithms, which produced an unexpected and unintended price escalation.
An accidental feedback loop between two automated retail pricing algorithms on Amazon inflated a textbook price to over 23 million dollars before manual intervention. The incident represents a classic example of algorithmic instability but has negligible national security implications.
The incident was driven by the dynamic interaction of two distinct algorithmic pricing agents adjusting their prices in response to each other, creating a collective feedback loop that neither agent would have exhibited alone.
Minimal or No Risk: The AI systems involved are simple algorithmic pricing tools used by third-party e-commerce sellers, which pose low or negligible risks to safety or fundamental rights.
The algorithms were designed to automatically determine and set the retail price of a book on Amazon based on competitor pricing.
The reports provide a clear, consistent, and detailed mathematical explanation of how the two pricing algorithms interacted to cause the price spiral. There is no conflicting information regarding the mechanics of the glitch.
None
Technology
Feedback Loop in Pricing Algorithms
The interaction of two algorithms created an upward price spiral.
Lack of Algorithmic Sanity Checks
Algorithms did not have built-in price ceilings or sanity checks.
Data Inputs
Competitor Price as Sole Input
The algorithms relied solely on the competitor price to set their own.
Human Factors
Lack of Active Human Oversight
Retailers failed to monitor the astronomical price increases for days.
Process and Methods
Flawed Arbitrage Pricing Strategy
Bordeebook priced at 1.27x competitor price to cover dropshipping costs.
Flawed Undercutting Strategy
Profnath consistently priced at 0.9983x competitor price to undercut.
Management
Inadequate Algorithmic Risk Assessment
Management failed to assess risks of automated pricing interactions.
AI system involved
Identified by automated extraction; can differ from the alleged developers above.
