AMLUCS 2026

Brought to you by:

Frazer-Nash Consultancy, partners of AMLUCS 2026
Frazer-Nash Consultancy, partners of AMLUCS 2026
Supporters AMLUCS 2025
Supporters AMLUCS 2025

Platinum Sponsors

Platinum Sponsor AMLUCS 2026
Platinum Sponsor AMLUCS 2026

Gold Sponsors

Gold Sponsors AMLUCS 2026
Gold Sponsors AMLUCS 2026
Gold Sponsor AMLUCS 2026

Venue Sponsor

Venue Sponsor AMLUCS 2026

Supported by

Supporters AMLUCS 2025
Supporters AMLUCS 2025
Supporters AMLUCS 2025
Supporters AMLUCS 2025
Supporters AMLUCS 2025
Supporters AMLUCS 2025

AMLUCS 2026

🎉What an incredible couple of days at Plexal, Here East, Queen Elizabeth Olympic Park for the fourth annual Applied Machine Learning for Cyber Security (AMLUCS) Conference.

We set out to create an event focused on the practical intersection of AI/ML and cyber security, bringing together the people researching, building, deploying and securing these technologies. ...Look what we've all achieved!

It was fantastic to see that ambition reflected in a sold-out conference and a community spanning central government, defence, major primes, SMEs and academia.

A few highlights:

  • A near 50/50 audience split across our two tracks, demonstrating the appetite for both Real-World & Applied perspectives and Research & Innovation.

  • A programme tackling some of the biggest questions facing the community, from agentic AI, autonomous cyber defence and AI security assurance, through to vulnerabilities, regulation, critical infrastructure and the security of the wider AI technology stack.

  • Talks that went beyond the technology itself, exploring the human element of AI security, including how we integrate autonomous systems with human judgement, oversight and accountability.

  • A continued focus on diversity, inclusion and the future pipeline of talent. Through our partnerships and bursary scheme, we were delighted to support early-career practitioners and academics to participate, learn, build their networks and contribute to the community.

But perhaps the biggest highlight was the community feel throughout.

AMLUCS is about bringing together different backgrounds, disciplines, organisations and perspectives, creating an environment where people can ask questions, share knowledge, challenge ideas and learn from one another. Seeing those conversations and new connections happening across government, industry and academia was hugely rewarding.

Selling out the conference was a fantastic milestone, but seeing the community grow, collaborate and share ideas is the real measure of success.

There is a huge amount happening at the intersection of AI and cyber security. The challenge now is to keep those conversations, connections and collaborations going long after the conference doors close.

See you at AMLUCS 2027! 🚀

Day 1

Real-World & Applied Track

Thematic Review and Gap Analysis of AI Security​​

Edward Austin, Lancaster University​

Panel: Where Do We Start? The Difference Between AI Security and Traditional Cyber Security

Ed Venmore-Rowland (Head of AI Cyber Security Policy, Department for Culture, Media and Sport)​
Harriet Farlow (Founder of Mileva Security Labs, AI Security researcher, author of 'Practical AI Security’)​
Dan Jones (Founder of Stealth Agentic AI Security, former Microsoft​ AI Red Team).​

Developing and Delivering the Australian Government’s First AI Security Framework Outside of National Security​

Harriet Farlow, ​Mileva Security Labs​

​Protecting the AI from the Human​​

Heather Taylor, ​Frazer-Nash Consultancy

Reproducing Mythos Findings with Public LLMs & Disclosing Two New Linux Kernel Zero-Days​

Klaudia Kloc, ​Vidoc Security Lab​

The Agentic Attack Surface: Latent Failures in Autonomous Workflows​

Ihsen Alouani, ​Queen’s University Belfast​

Autonomous Cyber Agents Meet Critical Infrastructure: Emergent Risks and Capabilities​

Chris Hicks, ​The Alan Turing Institute​

Integrating Agentic AI into a Human Decision Making Team​

Heather Taylor, ​Frazer-Nash Consultancy

Research & Innovation Track

Detecting Attacks Against AI Systems​

Louise Axon, University of Oxford​

Two Graders for One Agent​

Gilad Cohen, Alice​

Do Model Representations Know When a Jailbreak is Working?​

Sabrina Sadiekh, HiveTrace​

Latent Backdoors​

Katy Harrington, Loughborough University​

Hessian Spectral Signatures for Backdoor Detection​

Steven Young, The Alan Turing Institute​

Packets as Binary Images: Per-Packet Intrusion Detection for OT​

Aamir Omar, BMT​

Capture the Flag, Miss the Malware: Hijacking Agentic Code Scanners with Planted Rewards​

Gilad Cohen, Alice​

Planning for Quantum Backed AI​

John Sotiropoulos, DeepCyber and Roy Barker, Nebius​​

Agentic Large Language Models for Autonomous Cyber Operations​

Gregory Palmer, BAE Systems Digital Intelligence​

Day 2

Real-World & Applied Track

Sustainable Cyber Security in the Era of Hyperconnected Agentic AI and Cyber-Physical Systems​

Ryan Heartfield, Palo Alto​

Agentic AI for Cyber Security - An NCSC Perspective​

Paul J, NCSC​

Measured Feasibility of Agentic Cyber Defence at the Land Tactical Edge​

Ian Griffiths, Aeris UK​

Panel: Secure Enough? How Regulation Shapes AI Security​

Ali Shepherd (CISO and Chief Scientific Advisor, FCA),​
Dr. Alec Banks (Principal Advisor, Autonomous Systems & ​AI Safety Engineering, Dstl) ​
John Sotiropoulos (Founder of DeepCyber, Board Director of the OWASP GenAI Security Project, best selling AI Security author)​​

Investigating Agentic Cybercrime: a Forensic Model, its​
Judicial Validation, and a Reference Implementation​

John Sotiropoulos, DeepCyber and Marinos Bekios, Hellenic Police​

Agentic Security Benchmarks are Executable Threat Models: Measuring Risk Relocation in AI Defences​

Paul Vautravers, Advai​

Reasoning, Not Rationalising: Mechanistic Guardrails for Agentic AI in Critical National Infrastructure​

Don Liyanage and Mike Oaten, Tikos Technologies​

Automated Experimentation in Cyber Labs​

Alfie Beard, BT​

Research & Innovation Track

Behavioural Assurance of Agentic AI Systems in High-Stakes Environments​

Rick Hennessy, The Alan Turing Institute​

Identifiable Causal Discovery with Latent Time-varying Confounder Estimation (LaTICE)​

Susan Babirye, Loughborough University​

Security of Novel Compute in the AI Stack​

Sara Farmer, Dstl

​Hands-on Workshop - The Agent's Inner Monologue: Catching Compromise in the Reasoning Trace​

Most agent defences operate at the edges: input filters, guardrails, output classifiers and system prompts. But what if the most valuable security signal lies inside the agent itself? Inspired by recent advances in monitoring agent reasoning processes, this highly practical workshop explores how observing an agent's reasoning trace can reveal attacks and behaviours that action-level monitoring alone may miss. Bring your laptops and gain hands-on experience using Adrian, an open-source agent runtime monitoring platform.​

Max Corbridge & Sunny Chau, Secure Agentics ​