Will a production AI agent suffer a publicly disclosed prompt-injection breach before 2027?

As enterprises deploy autonomous AI agents with tool access, prompt injection is the leading systemic risk. This market tracks whether a real-world agent is compromised via prompt injection with public disclosure.

Resolution Criteria

Resolves YES if, before Jan 1 2027, a vendor advisory, incident report, or coverage from 2+ reputable outlets confirms a production/customer-facing AI agent was manipulated via prompt injection to exfiltrate data, take unauthorized action, or bypass controls. Resolves NO otherwise.