The Bank of Japan’s Institute for Monetary and Economic Studies has announced an online seminar for November 30 on security challenges arising from artificial intelligence (AI) agents. The event will examine risks associated with autonomous systems that act on users’ behalf, including operational failures, external attacks and vulnerabilities in web systems designed for human users. AI agents can misinterpret instructions, perform actions beyond their authorization or fail to verify results. External attacks can also cause unauthorized manipulation or data breaches, while autonomous access to web services may expose risks not addressed by conventional security designs. The seminar will explore security mechanisms that enable agents to interpret instructions correctly, remain within permitted boundaries and validate results without human involvement. Presentations will cover research on systems involving interacting AI agents and practical approaches to web security design and operation, drawing on observed attacks.
Bank of Japan announces November 30 seminar on security risks from AI agents
The Bank of Japan’s Institute for Monetary and Economic Studies will hold an online seminar on November 30 examining security risks from artificial intelligence (AI) agents. Topics include misinterpreted instructions, actions beyond authorized limits, inadequate verification and external attacks. The sessions will address safeguards for autonomous agents and changes to web security design and operation needed to accommodate their access.