According to BeInCrypto, Daniel Ishag, chief commercial officer at Bakkt, described a hypothetical scenario in which an AI agent books a business-class flight to Korea that fits a user’s budget but selects the wrong destination among multiple locations named Korea. The example was given during a panel discussion and serves to illustrate potential issues when AI agents make travel-related decisions on behalf of users.
The scenario highlights a concern about AI agents choosing options that appear valid within set parameters—such as price range and class level—but result in incorrect destinations due to ambiguous or similarly named locations. No specific flight, airline, date, user name, or numerical details were provided. The discussion focused on the need to clarify responsibility when an agent’s choice does not meet the user’s intended destination.
No official policy, regulation, or legal determination was mentioned regarding who would be responsible—such as the user, the agent developer, or another party—if an AI agent books a flight to the incorrect Korea. The remarks were presented as illustrative of emerging challenges in real-world use of AI agents for booking decisions.
