Cybersecurity in the Era of LLMs: Governance, Runtime Protection, and Automation

Hosted by Tokyo AI in collaboration with the Japan Institute for Cyberspace Studies (JICSS), this session examines the technical intersection of AI and cybersecurity. It opens with a panel on the evolving zero-trust environment and emerging compliance mandates, bringing together senior policy practitioners and frontline engineers to bridge AI governance, international compliance, and secure software architecture. Panelists include Ilya Kulyatin (Tokyo AI, Foundry Labs), Yury Leonychev (Wallarm), James Miller (Miller Enterprises), and Dr. Anand R. Prasad (Deloitte). After an hour of networking, three expert talks dig into securing AI deployments. Ramy Aouinet (Antitech) covers the broad threat landscape for LLMs and the need for unified, end-to-end security infrastructure. Yury Leonychev (Wallarm) addresses runtime discovery, observability, and inline enforcement for agentic AI in production. Arpit Jain (ex-KPMG Japan) closes with automated vulnerability remediation across large-scale open-source repositories, drawing on over 150 merged security pull requests in projects like Kubernetes, NumPy, and Apache Kafka. The format differs from regular TAI events: it starts earlier with the panel, then networking, then the talks. Attendees can join either or both parts. The event is held at CIC Japan in Minato City.