Abstract: As 6G networks become increasingly autonomous and AI-driven, ensuring their security, trustworthiness and accountability presents new challenges. A security-by-design approach to Agentic AI offers a way to address these challenges by embedding machine-readable security requirements into autonomous decision-making processes. We explore how standardised records of AI-agent decisions can improve transparency, accountability and provenance, making it possible to understand and trace the actions of autonomous agents. We then present two complementary contributions towards enabling autonomous yet explainable Radio Access Network (RAN) security, combining security requirements with trustworthy decision-making. Together, these approaches lay the groundwork for more secure, transparent and accountable AI-driven networks in the 6G era.
Date: Wednesday 28 October 2026
Time: 12:30–16:15
Venue: Dorothy Fleming Lecture Theatre, 133 Charles Street, Sheffield, S1 2ND
Agenda
12:30–13:00: Arrival, light lunch, refreshments and networking.
13:00–14:00: Keynote by Dr Mohammad Shojafar — Securing the agentic workforce: from discovery to control.
14:00–14:20: Break and networking.
14:20–15:00: Panel discussion chaired by Warren Atkinson.
15:00–15:20: Break and networking.
15:20–16:00: Keynote by Scott Knowles — Beyond Infrastructure: Securing the Intelligence of Tomorrow.
16:00–16:15: Closing remarks and thanks.
Sponsored by CRANE, the EPSRC NetworkPlus for the cyber security research community