What does an attacker look for when your company has AI agents everywhere?

In this special episode of Radio Logic, we sit down with OTW (Occupy The Web), founder of Hackers Arise. OTW has trained US military and intelligence teams in hacking and cyber security, and has run the Hackers Arise training programme for over 10 years.

We talk about:

• Why your data should stay local, not with remote AI models
• Why a human must stay in the loop for key security decisions
• How attackers probe AI agents to find their weak spots
• The risk of forgotten "rogue" agents holding live keys and tokens
• Why open source models are catching up with the big players

We also dig into the Wittgenstein Tournament, a Hackers Arise contest to build the best defensive agentic AI. Over 200 people have already signed up, and the deadline has been pushed back to 31 October.

Want to take part? Head to the Hackers Arise website to sign up: https://hackers-arise.com/wittgenstein-ai-tournament/

#CyberSecurity #AgenticAI #IdentitySecurity #RadioLogic #RadiantLogic #AIsecurity #HackersArise

Breaking Into Cyber: A 20-Year-Old's Guide to Ethical Hacking and Identity Security

What if the best way to defend a business is to learn how to break into it?

In this episode of Radio Logic, we sit down with Samuel, a 20-year-old placement student from Coventry University, one of the only UK universities teaching ethical hacking. Samuel shares how he found cyber during lockdown, why he chose the attacker's mindset, and what it's really like rotating through pen testing, identity, SOC, and governance teams at FSP.

We dig into:

• Why cyber is a huge field, not one job
• The 4.8 million person gap in cyber talent and what it means for newcomers
• How AI and deepfakes are reshaping social engineering threats
• Why your social posts could put your career at risk
• Free routes into cyber (no degree required): Try Hack Me, Hack The Box, and more

Samuel's big message? Stay curious. The exploit might take months to find, but the door is open to anyone willing to poke around.

New to identity and security? This one's for you.

Your network is locked down. Your endpoints are secure. So why are you still getting breached?

Anders Åskåsen sits down with Simon Moffett of The Cyber Hut to make the case that identity is the real control plane of security.

They dig into why identity became a single point of failure, why logs only tell you what already went wrong, and why "just-in-time" access is killing the old pre-provisioned model.

Simon's takeaway: you can have the fanciest AI in the world, but if you can't log in, none of it works.

For anyone who owns security, identity, or risk. Press play.

How does your body fight off millions of threats it has never seen before, without a single rulebook?

That question might hold the key to securing AI agents.

Anders Åskåsen sits down with Tom from FSP to explore a fresh way of thinking about identity: borrow from biology.

They dig into why role-based access keeps exploding, why we never solved least privilege even for simple service accounts, and why agents now run those same old problems millions of times faster.

Tom's big idea? Treat security like an immune system. Block the known-bad on sight. Slow down the suspicious. Give a human time to step in.

Plus the wild story of an AI agent that signed up for its own API token overnight and would not stop making phone calls.

If you own identity, security, or governance, press play.

Would you give one new hire 45 times more accounts and 100 times more data, then never check on them?

That is what businesses are doing with AI agents right now.

Anders Åskåsen sits down with Michael from Cyber IAM to dig into the messy reality of governing Agentic AI.

They cover why so many AI projects stall at the finish line, why every business now has a "Shadow AI" problem, and why the fix might be a gateway with real guardrails, not blind trust.

Michael's rule is simple: it is called a copilot for a reason. You are still flying the plane.

If you own identity, security, or governance, press play.

What happens to identity security when AI agents start making their own decisions?

Anders Åskåsen sits down with Simon Moffett of The Cyber Hut to unpack the collision between Agentic AI and identity governance.

They cover why IGA and PAM projects keep failing, what makes AI agents different from normal machine identities, and why "Shadow AI" could be your biggest blind spot.

Plus the line you will not forget: would you give a four-year-old the keys to a Ferrari?

If you own identity, security, or governance, press play.

🔐 Identity used to sit behind a firewall. Now it IS the perimeter.

In this episode of Radio Logic, Anders Åskåsen sits down with Simon Moffett, founder and analyst at The CyberHut, to unpack one of the biggest hidden costs in enterprise tech: identity debt. 💸

They dig into why old identity choices come back to bite, how shadow IT became shadow identity, why moving to the cloud didn't fix the mess, and where smart CISOs are moving their budget. If your identity layer is slowing the business down, this one is for you. 🎧

How long would it take to fake you? Your voice, your face, the way you write. According to this episode's guest, about one hour. By next year, maybe fifteen minutes, maybe one minute.

In this episode of Radio Logic, the show that breaks identity down into something you can actually use, host Anders Askasen sits down with Matt Harris to unpack how fast AI voice and face cloning has moved, and what it means for trust, security, and the people we most want to protect.

They cover how a convincing clone now needs just one minute of audio and a few photos, the $25 million deepfake video call that fooled a finance team, why the old tell-tale signs like six fingers are gone, the gap left by AI tools with no guardrails, and the simple moves that cut your risk today.

Identity is the number one cause in every breach report. So why haven't we fixed it?

In this episode of Radio Logic, the show that breaks identity down into something you can actually use, host Anders Askåsen sits down with Tom Hebbron, Head of Identity and Security Architecture at FSP, to unpack "identity debt", why it keeps growing, and how to start clawing it back.

They cover why identity data ends up siloed across the business, the speeds at which access decisions have to happen (milliseconds at the edge versus months in forensics), what the Verizon and IBM breach reports keep telling us, why annual access reviews leave data 11 months stale, why no access should ever be handed out without an end date, and how dynamic, context-aware access beats the role-based access trap.

Why do so many identity projects fail? And how do the best ones actually get done?

In this episode of Radio Logic, the show that breaks identity down into something you can actually use, host Anders Askasen sits down with Michael Ribaudo from CyberIAM to unpack why identity projects stall, drag, or collapse, and what separates the ones that succeed.

They cover why identity work never really ends, why an outcome-based approach beats a feature list, why the CEO (not IT) should sponsor the project, the trouble with role-based access when a 60,000-person bank ends up with 350,000 roles, the shift to policy-based and continuous access, and the honest advice Michael gives any customer about to scope a new project.