AI is advancing at a pace that leaves even the most forward‑thinking organisations struggling to keep up. While academia pushes the boundaries of what’s possible, industry leaders are left trying to translate that innovation into practical resilience, often without the time or access to understand the latest breakthroughs.
That’s exactly why the CRANE AI-Cyber Special Interest Group (SIG) exists.
CRANE (Cyber Security Research and Networking Environment) is a UK-wide research network funded by UKRI to support the delivery of the National Cyber Strategy and strengthen the UK’s cyber security research ecosystem. The quarterly AI-Cyber SIG aims to connect industry with the latest research shaping the future of AI, helping them prepare for the challenges of tomorrow.
Last week’s inaugural AI-Cyber SIG brought together over 50 professionals from academia and industry to see 24 presentations of cutting-edge research from leading UK universities.
The day kicked off with the keynote address from the NCSC’s Ollie Whitehouse, who delivered a sharp wake-up call: as AI scales the speed and sophistication of cyber threats, traditional “guardrails” are no longer enough. To survive a machine-speed threat environment, the UK must shift toward secure-by-design technology and proactive, board-level accountability.
Below are just a few small highlights from a packed agenda…
Securing AI and preventing threats
Agentic AI Transformation: Agentic AI is transforming cyber. While the road may be rocky, AI will ultimately be a net positive for cyber security strengths. However, this new paradigm raises immediate security issues, such as establishing proper access management and payment permissions.
Advanced Threat Detection: Researchers across the UK covered a range of practical defense strategies, including LOTL-Hunter for detecting multi-stage living-off-the-land (LOTL) attacks in cyber-physical systems (CPS), and Agentic SABRE, an uncertainty-aware multi-agent framework for adaptive ransomware detection.
Vulnerability in Robotics and LLMs: Collaborative Robots (Cobots) have security problems. They’re fantastic resources but they’re vulnerable to insider attacks, and there is a potential for hacking through Prompt Injection Attacks on LLMs. Don Munasinghe from the University of West London discussed the significant impact semi-autonomous real-time Intrusion Detection and Prevention Systems (IDPS) can have on industries using cobotics such as manufacturing, healthcare and automotive.
Combating Disinformation and Building Trust
AI is increasingly being weaponised against human cognition. As trust becomes harder to maintain, research into safeguarding information ecosystems is more critical than ever.
Cognitive Warfare: Research from the University of Strathclyde highlighted the rapid evolution of AI-driven content, which is creating a new age of cognitive warfare through scalable disinformation, deepfakes, and microtargeted influence campaigns. This targets perception, trust, and decision-making.
Trustworthy AI: CRANE’s AI researchers addressed the ethical risks of unsupervised LLMs. Academics are making huge progress, and the future of trustworthy AI is exciting. One example that showed the bright future ahead was Sheffield Hallam University’s work with the NHS to ensure a LLM does not give unacceptable advice, lack empathy, or expose patient information.
Resilience to Fake News: Fake News travels six times faster than True News. It’s a terrifying fact, but researchers are working hard to combat this. Models such as a Forensically Informed Multi-Model Machine Learning Framework for Fake News Detection were discussed in detail.
These topics, among many others, were explored in depth by the researchers who are leading the charge in this rapidly evolving field.
Join the Next Conversation.
The CRANE AI-Cyber quarterly SIG gives industry professionals exclusive access to the latest research that is shaping the next generation of security and resilience.
By joining, industry professionals gain:
- Direct access to leading AI researchers
- Early insight into emerging threats and defensive technologies
- A clearer understanding of how academic innovation can be applied in real‑world environments
If your organisation is committed to staying resilient and informed in a rapidly evolving AI landscape, becoming a member of the SIG is an opportunity you cannot afford to miss. Join here.
In the meantime, we’ll be keeping you up to date with all of the latest developments over on the CyberNorth LinkedIn.
Stay tuned for the next one!
