ABOUT THIS EPISODE
You are terrified of machines taking over, yet you wake up every morning and operate your schedule like one.
In our collective rush to automate heavy lifting, enterprises are blindly handing the keys to critical infrastructure over to autonomous AI agents operating with zero human visibility. The "Move Fast and Break Things" era is dead. Today, an unvetted agent given a completely benign prompt- combined with access to sensitive data and external networks- can quietly misalign, bypass permissions, and infiltrate live production. The ultimate leadership test is no longer about how fast you can deploy AI, but whether you have the architecture to actually see what it is doing and govern it.
Live from the floor at the International Cyber Expo (ICE) at Olympia London 2026, we sit down with Jack Miller, Co-Founder & CTO of Helmguard AI. Moving past generative AI hype, we expose the real mechanics of agentic AI security, why elite enterprises choose 16-hour governed workflows over 1-hour unmonitored scripts, and how to deploy secure execution environments to keep autonomous tech in check.
A massive thank you to Helmguard AI for partnering with us to sponsor this on-site ICE special edition episode.
🔥 Inside this episode:
- The Misalignment Reality: How simple prompts trigger rogue network actions and the growing legal threat of autonomous agent liability.
- The 16-Hour Rule & GRC 8.0: Why top hedge funds prioritize slow, rigorous assurance and how to simulate risk before breaking production.
- Taming the AI: Fixing LLM sycophancy, eliminating shadow AI, and safely automating vendor governance.
- Scaling Securely: How Helmguard AI reached £1M ARR in 9 months by cutting through the noise to master AI security.
🚀 Featured Guest & Sponsors:
Jack Miller, HelmGuard AI’s CTO, is an AI engineer and researcher specializing in secure, robust AI systems. He has collaborated with top European institutions such as ETH Zürich, University of Cambridge, and University of Oxford. Jack graduated from the Australian National University, earning one of the country’s most prestigious undergraduate scholarships. His work includes publishing in leading AI research journals and developing a commercialized algorithm in quantum chemistry. Jack combines deep academic expertise with practical innovation to shape HelmGuard’s approach to AI-powered cyber defence.
HelmGuard is currently offering a free rapid assessment of your company's AI usage and risks. We will review up to 50 vendors you select to determine the risk posed by their use of AI. We also check your internal AI usage via a risk assessment process that can bring in data from MDM, internal policies and your cloud usage.
👉 Get in touch: https://helmguard.ai/contact
Up Next: SUBSCRIBE so you don't miss our upcoming montage episode from the ICE floor featuring rapid-fire operator insights. Special thanks to Helmguard AI for making this series possible.
The £1M Enterprise AI Challenge: Through SECURE | CYBER CONNECT, we are deploying £100,000 in fully funded, hands-on cybersecurity and AI infrastructure support to TEN selected individuals, leaders, and nonprofits.
🎟️ Register Interest: https://docs.google.com/forms/d/e/1FAIpQLScdcQpsUsBp50krYU_IW3rMcuw-Kk5RYATMp5lCA1vfwvObEA/viewform
Connect On LinkedIn:
• Warren: https://www.linkedin.com/in/warren-atkinson/
• Jay: https://www.linkedin.com/in/justinadamson/
• Jack: https://www.linkedin.com/in/jack-w-miller/
Join Us Next Year at ICE: https://www.internationalcyberexpo.com/
📲 Free guides, tools and resources: https://link.v1ce.co/warrenatkinson
🔔 Please SUBSCRIBE, LIKE + SHARE to empower others.
In our collective rush to automate heavy lifting, enterprises are blindly handing the keys to critical infrastructure over to autonomous AI agents operating with zero human visibility. The "Move Fast and Break Things" era is dead. Today, an unvetted agent given a completely benign prompt- combined with access to sensitive data and external networks- can quietly misalign, bypass permissions, and infiltrate live production. The ultimate leadership test is no longer about how fast you can deploy AI, but whether you have the architecture to actually see what it is doing and govern it.
Live from the floor at the International Cyber Expo (ICE) at Olympia London 2026, we sit down with Jack Miller, Co-Founder & CTO of Helmguard AI. Moving past generative AI hype, we expose the real mechanics of agentic AI security, why elite enterprises choose 16-hour governed workflows over 1-hour unmonitored scripts, and how to deploy secure execution environments to keep autonomous tech in check.
A massive thank you to Helmguard AI for partnering with us to sponsor this on-site ICE special edition episode.
🔥 Inside this episode:
- The Misalignment Reality: How simple prompts trigger rogue network actions and the growing legal threat of autonomous agent liability.
- The 16-Hour Rule & GRC 8.0: Why top hedge funds prioritize slow, rigorous assurance and how to simulate risk before breaking production.
- Taming the AI: Fixing LLM sycophancy, eliminating shadow AI, and safely automating vendor governance.
- Scaling Securely: How Helmguard AI reached £1M ARR in 9 months by cutting through the noise to master AI security.
🚀 Featured Guest & Sponsors:
Jack Miller, HelmGuard AI’s CTO, is an AI engineer and researcher specializing in secure, robust AI systems. He has collaborated with top European institutions such as ETH Zürich, University of Cambridge, and University of Oxford. Jack graduated from the Australian National University, earning one of the country’s most prestigious undergraduate scholarships. His work includes publishing in leading AI research journals and developing a commercialized algorithm in quantum chemistry. Jack combines deep academic expertise with practical innovation to shape HelmGuard’s approach to AI-powered cyber defence.
HelmGuard is currently offering a free rapid assessment of your company's AI usage and risks. We will review up to 50 vendors you select to determine the risk posed by their use of AI. We also check your internal AI usage via a risk assessment process that can bring in data from MDM, internal policies and your cloud usage.
👉 Get in touch: https://helmguard.ai/contact
Up Next: SUBSCRIBE so you don't miss our upcoming montage episode from the ICE floor featuring rapid-fire operator insights. Special thanks to Helmguard AI for making this series possible.
The £1M Enterprise AI Challenge: Through SECURE | CYBER CONNECT, we are deploying £100,000 in fully funded, hands-on cybersecurity and AI infrastructure support to TEN selected individuals, leaders, and nonprofits.
🎟️ Register Interest: https://docs.google.com/forms/d/e/1FAIpQLScdcQpsUsBp50krYU_IW3rMcuw-Kk5RYATMp5lCA1vfwvObEA/viewform
Connect On LinkedIn:
• Warren: https://www.linkedin.com/in/warren-atkinson/
• Jay: https://www.linkedin.com/in/justinadamson/
• Jack: https://www.linkedin.com/in/jack-w-miller/
Join Us Next Year at ICE: https://www.internationalcyberexpo.com/
📲 Free guides, tools and resources: https://link.v1ce.co/warrenatkinson
🔔 Please SUBSCRIBE, LIKE + SHARE to empower others.
English
United States
TRANSCRIPT 🔗
Are you the producer of this podcast?
Add a podcast transcript
Need Audio-to-Text?
Transcribe with Listen411 in Just 60 Seconds
SEARCH PAST EPISODES
Search past episodes of CYBER CONNECT PODCAST.
OTHER EPISODES IN THIS PODCAST
Let's be real about the state of traditional networking.
Leaders are completely exhausted by the endless cycle of vendor pitches.
Getting dragged from one surface-level meeting to the next does not solve real problems.
Jay Adamson and Warren Atkinson sat down with Llewellyn James and Jack Arms…
Searching for answers on why your enterprise sandbox keeps breaking? Want to know how to deploy production-grade trust infrastructure cleanly? You are in the exact right place.
Here is the reality: You build a brilliant AI tool in a safe testing space and it works perfectly. But the second you pl…
If you work hard, follow all the traditional rules, and keep your head down, why do you still feel overlooked, exhausted, and stuck in your career?
We went behind the scenes at the CyBlack Conference 2026 to capture raw, unfiltered stories from engineers, keynote speakers, and frontline builders …
Forget the Hollywood myths of green matrix code and black hoodies. The reality of how your company gets hacked is a clinical, psychological, and highly industrialised operation.
In this episode, Jay Adamson and Warren Atkinson are joined by Ethical Hacker Glenn Wilkinson lifts the veil on the mur…
Let’s have an honest conversation about the cybersecurity job market.
You have put in the hard miles. You have studied late into the night, paid for the certifications, and sent endless applications into the void- only to hit a brick wall. If you are feeling exhausted, isolated, or wondering if yo…
Disclaimer: The podcast and artwork embedded on this page are from Jay Adamson & Warren Atkinson, which is the property of its owner and not affiliated with or endorsed by Listen Notes, Inc.
EDIT
Thank you for helping to keep the podcast database up to date.
