{"data":{"id":"3cdc863a-117f-4e94-ab09-cf68060a6086","title":"The Race Is on to Keep AI Agents From Running Wild With Your Credit Cards","summary":"Agentic AI (AI systems that perform actions on behalf of humans) is growing in use, but it creates new security risks like agents being hijacked or tricked into unauthorized transactions. The FIDO Alliance (an industry group focused on authentication standards), along with Google and Mastercard, is launching working groups to develop security standards that will protect AI agent transactions using cryptographic tools (mathematical techniques that verify identity and prevent tampering) and authentication mechanisms that prevent phishing attacks.","solution":"Google is contributing the Agent Payments Protocol (AP2), which cryptographically verifies that a user intended for an agent-initiated transaction to happen. Mastercard is contributing the Verifiable Intent framework (codeveloped with Google), which is a secure mechanism for users to authorize and control agent actions. Together, these tools aim to provide cryptographic proof that transactions were authorized by the user while maintaining privacy through selective disclosure, so different parties in the payment ecosystem only see relevant information.","labels":["policy","security"],"sourceUrl":"https://www.wired.com/story/the-race-is-on-to-keep-ai-agents-from-running-wild-with-your-credit-cards/","publishedAt":"2026-04-28T13:00:00.000Z","cveId":null,"cweIds":null,"cvssScore":null,"cvssSeverity":null,"severity":"info","attackType":[],"issueType":"news","affectedPackages":null,"affectedVendors":["Google","Microsoft"],"affectedVendorsRaw":["Google","Mastercard","FIDO Alliance"],"classifierModel":"claude-haiku-4-5-20251001","classifierPromptVersion":"v3","cvssVector":null,"attackVector":null,"attackComplexity":null,"privilegesRequired":null,"userInteraction":null,"exploitMaturity":null,"epssScore":null,"patchAvailable":null,"disclosureDate":"2026-04-28T13:00:00.000Z","capecIds":null,"crossRefCount":0,"attackSophistication":"moderate","impactType":["integrity","availability"],"aiComponentTargeted":"agent","llmSpecific":false,"classifierConfidence":0.85,"researchCategory":null,"atlasIds":null}}