Kalshi develops AI agent "Harrison" to stress-test prediction market contracts and betting logic

Odaily reports that prediction market platform Kalshi is internally deploying an AI agent system named "Harrison" to optimize its prediction market contract designs and stress-test betting outcomes, aiming to reduce the risk of errors and ambiguities in large-scale trading. According to co-founder Luana Lopes Lara, the AI tool is being used to handle critical internal processes for the platform's millions of daily transactions, focusing on resolving complexities in the wording and rules of event contracts covering areas such as political elections, sports events, and award outcomes.

Kalshi stated that "Harrison" is helping the team identify potential vulnerabilities in contract designs in advance, reducing disputes caused by unclear definitions, thereby enhancing market operational stability and execution efficiency. As the scale of prediction market trading expands, this AI system is seen as a significant upgrade to the infrastructure layer, contributing to improved contract standardization and automated governance capabilities. (Bloomberg)