AI Security, AI Cloud Ops & Infrastructure | ClearTech Loop

What happens when AI becomes capable of making decisions its creators can't fully anticipate or control?

AI agents are gaining autonomy, accessing enterprise systems, managing credentials, and making decisions at speeds traditional security frameworks weren't designed to handle. But as these systems become more intelligent, are organizations becoming better at governing them?

In this episode of ClearTech Loop, Jo Peterson sits down with Sandy Dunn, CISO at Knowtion Health and a core member of the OWASP Top 10 for LLM Applications project, to explore what AI governance really means when intelligence, autonomy, and accountability collide.

Sandy compares governing advanced AI to raising a child who's smarter than you. It's a provocative analogy that raises an important question: How do you establish boundaries for a system that may be capable of finding solutions you never anticipated?

Their conversation moves beyond traditional cybersecurity controls to examine AI agent identity, least privilege, unpredictable behavior, business risk, and the growing responsibility of security leaders to enable innovation without surrendering oversight.

In This Episode

• Why traditional identity and access management approaches fall short for autonomous AI agents
• How unique agent identities improve accountability and traceability
• Why increasingly intelligent AI systems don't necessarily demonstrate good judgment
• What happens when agents pursue objectives in unexpected or unauthorized ways
• How CISOs should approach organizations already behind on AI governance
• Why security controls must support business innovation rather than obstruct it
• How executive education and experimentation influence successful AI adoption
• Why understanding AI model costs, performance, and business value matters

Worth Reading + Watching

• AI Threat Map v2.1: Navigating AI Threats & Tradeoffs for AI Adoption Sandy Dunn examines how organizations can navigate AI security risks while supporting innovation. https://www.linkedin.com/pulse/ai-threat-map-v21-navigating-threats-tradeoffs-adoption-sandy-dunn-zprgc
• OWASP Top 10 for LLM Applications Explore vulnerabilities and security risks associated with large language model applications, including excessive agency and prompt injection. https://genai.owasp.org/llm-top-10/
• ClearTech Loop with Rob Tiffany: AI Agent Identity Continue the conversation about agent identities, authorization, and why traditional identity management may not be sufficient for autonomous systems. https://www.youtube.com/watch?v=4Lr4OVcBjEo

Stay in the Loop

Subscribe to Jo Peterson's LinkedIn newsletter for more conversations about AI security, cyber risk, and enterprise technology:

https://www.linkedin.com/newsletters/7346174860760416256/

Subscribe to ClearTech Loop on YouTube:

https://www.youtube.com/channel/UCR7xPyjXPgQJKMBvnjoGBQw

ClearTech Loop is where AI, cloud, and security intersect. New conversations with the leaders shaping what's next in enterprise technology.

See you in the Loop.

🎧 Listen: In Buzzsprout Player
▶ Watch on YouTube: https://www.youtube.com/@ClearTechResearch/videos
📰 Subscribe to the Newsletter:
https://www.linkedin.com/newsletters/7346174860760416256/  

Your company has an AI governance committee. Great. But when an AI agent makes a decision that harms your business, who actually has the authority to shut it down?

And here's a bigger question: Would you even know it made the wrong decision?

In this episode of ClearTech Loop, host Jo Peterson sits down with Maribel Lopez, Founder and Principal Analyst at Lopez Research, to examine the uncomfortable reality behind enterprise AI adoption.

Organizations are moving quickly to deploy AI, but many haven't established clear accountability, reliable data practices, or meaningful ways to measure whether their investments are actually improving the business.

Maribel challenges the assumption that forming a governance committee solves the problem. She explains why AI can make technically logical decisions that are completely wrong for the business, and why governance becomes even more critical as AI agents move from recommending actions to taking them.

In This Episode

• Who's actually in charge? Why AI governance committees can fail without someone empowered to intervene.
• When AI goes rogue: Why human oversight, permissioning, and rollback capabilities matter.
• The productivity trap: Why making someone's job 30% faster doesn't necessarily create business value.
• The business context problem: How AI can use accurate data and still make a disastrous decision.
• Data you can't trust: Why disconnected systems and conflicting information undermine AI outcomes.
• Decisions you can't explain: What happens when organizations must justify an AI-driven decision.
• The executive responsibility: Why AI governance extends beyond IT and security.

The question isn't whether AI can make decisions. It's whether your organization can stand behind them.

About Our Guest

Maribel Lopez is the Founder and Principal Analyst at Lopez Research and an enterprise AI analyst who advises organizations on emerging technologies, AI adoption, and business transformation.

Website: https://lopezresearch.com/

About Our Host

Jo Peterson is the CIO of Clarify360 and Chief Analyst at ClearTech Research. Through ClearTech Loop, she explores the technology trends, risks, and decisions shaping enterprise leadership.

Worth Reading + Watching

• AI Has a Business Context Problem | Maribel Lopez Why accurate data alone isn't enough for AI-driven business decisions. https://lopezresearch.com/2026/05/06/ai-has-a-business-context-problem/
• The Hard Part of Agentic AI Isn't Building the Agent. It's the Space Between Your Tools | Maribel Lopez The challenges of connecting AI systems while maintaining context, security, and governance. https://lopezresearch.com/2026/08/02/the-hard-part-of-agentic-ai-isnt-building-the-agent-its-the-space-between-your-tools/
• AI Agents Are Already Working for You. Do You Know Who They Are? | ClearTech Loop with Michael Krigsman A related conversation about visibility, ownership, and securing an AI workforce. https://cleartechresearch.com/ai-agent-security-michael-krigsman/

Stay in the Loop

For more conversations challenging the assumptions behind enterprise AI, cybersecurity, and emerging technology, follow ClearTech Loop.

Subscribe on YouTube:

https://www.youtube.com/channel/UCR7xPyjXPgQJKMBvnjoGBQw

Subscribe to Jo's LinkedIn newsletter:

https://www.linkedin.com/newsletters/7346174860760416256/

🎧 Listen: In Buzzsprout Player
▶ Watch on YouTube: https://www.youtube.com/@ClearTechResearch/videos
📰 Subscribe to the Newsletter:
https://www.linkedin.com/newsletters/7346174860760416256/  

Your company wouldn't hire an employee without onboarding, supervision, or defined access. So why are AI agents being given responsibilities without those same safeguards?

In this episode of ClearTech Loop, host Jo Peterson sits down with Rasha Horn, Founder & Principal Advisor at CyberNarwhal.ai, to examine what happens when organizations deploy an AI workforce faster than they can secure it.

Rasha makes the risks remarkably relatable. From agents creating other agents to governance committees that lack operational oversight, she challenges business leaders to rethink how they manage AI security and accountability.

In This Episode:

• Why AI agents need the equivalent of employee onboarding
• The hidden risks of autonomous agents and delegated access
• Why governance committees alone aren't enough
• Who should own, monitor, and be accountable for AI agents
• Why legacy infrastructure matters to AI security
• The question: Would you give an AI agent more freedom than a new employee?

Stay in the Loop. 
Subscribe on YouTube: https://www.youtube.com/@ClearTechResearch
Follow Jo's LinkedIn newsletter: https://www.linkedin.com/newsletters/7346174860760416256/

🎧 Listen: In Buzzsprout Player
▶ Watch on YouTube: https://www.youtube.com/@ClearTechResearch/videos
📰 Subscribe to the Newsletter:
https://www.linkedin.com/newsletters/7346174860760416256/  

AI agents are moving across systems, making decisions and, in some cases, creating other agents. But enterprises still do not have consistent standards for identifying them, governing their permissions or tracing what they do.

In this episode of ClearTech Loop, Jo Peterson talks with Rob Tiffany, Research Director at IDC, about why traditional identity and access management may not be enough for the agentic AI era.

Rob argues that the identity problem becomes much bigger once agents begin operating beyond a single enterprise. Companies need to know who an agent belongs to, what it is allowed to do and whether its activity can be audited and traced. As Rob puts it, there are still no consistent standards around agent “identity and permissions and auditability and traceability.”  

He also discusses an idea he is exploring with Cerf: whether AI agents may eventually need something similar to a global registry, analogous to DNS, so their identity, permissions and traceability can be verified across networks and organizations. 

The conversation also gets into shadow AI, employee adoption and why the most productive people inside an organization may be among the first to use AI tools before governance catches up. 

In This Episode

• Why traditional IAM was not built for autonomous AI agents
• How software agents have existed for decades, from scripts to RPA
• Why identity becomes harder when agents create other agents
• The need for permissions, auditability and traceability across organizations
• Why AI agents may eventually need a global identity layer
• How shadow AI is entering enterprises faster than governance can respond
• What CISOs should be thinking about when AI adoption is already ahead of controls
• Why productivity cannot come at the expense of enterprise security

🎧 Listen: In Buzzsprout Player
▶ Watch on YouTube: https://www.youtube.com/@ClearTechResearch/videos
📰 Subscribe to the Newsletter:
https://www.linkedin.com/newsletters/7346174860760416256/  

AI agents are already operating inside enterprises, accessing systems, making decisions and, in some cases, creating other agents.

The problem? Many organizations still cannot answer some very basic questions: What agents are running? Who owns them? What can they access? And who is accountable for what they do?

In this episode of ClearTech Loop, Jo Peterson talks with Michael Krigsman, founder and host of CXOTalk, about what AI security looks like when the workforce is no longer entirely human.

Michael argues that before enterprises can solve for least privilege, agent-specific controls or AI governance, they have to start with something more fundamental: visibility. Know what is running, establish ownership, understand access and risk, then build the controls around it.  

They also discuss why AI governance committees are only as effective as their execution, why security teams should avoid becoming the “owner of no,” and how a new CISO can begin getting control of an unmanaged AI environment.  

In this episode:

• Why traditional IAM and PAM were not built for autonomous AI agents
• Why agent inventory and visibility have to come first
• What happens when agents create subagents with their own permissions
• The gap between AI governance policy and actual enforcement
• How leaders should think about shadow AI
• Where a new CISO should begin in the first 30 days
• Why AI agents may need to be managed more like employees, contractors or freelancers

Michael leaves leaders with a practical way to think about the problem: give agents identities, know what jobs they are doing, know who owns them and know what they can access.  

About Michael Krigsman

Michael Krigsman is a globally recognized analyst, strategic advisor and industry commentator focused on digital transformation, innovation and AI leadership. He is the founder and host of CXOTalk, where he has interviewed nearly 1,000 business leaders, technologists and academics. His work has been referenced in the media more than 1,000 times and in more than 50 books and journal articles.  Pasted markdown

Worth Reading + Watching

Hugging Face: Anatomy of a Frontier Lab Agent Intrusion
A technical look at the incident discussed during the episode and how autonomous agent activity can move across systems and trust boundaries.

Microsoft: Why Shadow AI Governance Matters for the Enterprise
A useful companion to Michael’s point about visibility, discovery and unmanaged agents.

ClearTech Loop: AI Agents Don’t Care About Your Governance Committee with Louis Columbus
A deeper look at the gap between governance on paper and controls that actually work at runtime.

ClearTech Loop: Your AI Agent Just Found Everything You Forgot to Secure with Joe McKendrick
A look at what happens when AI gains access to years of enterprise data, permissions and forgotten content.

Stay in the Loop

Subscribe to ClearTech Loop for conversations with technology leaders about AI, cybersecurity, cloud and the issues changing enterprise technology right now.

And one question to take back to your team:

Could your organization produce a complete inventory of every AI agent operating today?

🎧 Listen: In Buzzsprout Player
▶ Watch on YouTube: https://www.youtube.com/@ClearTechResearch/videos
📰 Subscribe to the Newsletter:
https://www.linkedin.com/newsletters/7346174860760416256/  

AI agents are gaining credentials, permissions and the ability to act across enterprise systems. But what happens when they discover years of forgotten files, excessive permissions and data that was never secured with AI access in mind? 

In this episode of ClearTech Loop, Jo Peterson talks with longtime technology analyst and Forbes contributor Joe McKendrick about the security implications of the agentic workforce. 

They discuss: 

• Why AI agents should be treated like non-human identities 
• The gap between AI governance policies and actual enforcement 
• How agents can expose forgotten enterprise data and permissions 
• Why human oversight and accountability still matter 
• What security leaders should be thinking about as agentic AI expands 

Joe offers a simple warning: treat your AI agents with the same caution and restricted access you would give an intern. 

Because AI may not create your security debt. 

It may simply be the thing that finally exposes it. 

🎧 Listen: In Buzzsprout Player
▶ Watch on YouTube: https://www.youtube.com/@ClearTechResearch/videos
📰 Subscribe to the Newsletter:
https://www.linkedin.com/newsletters/7346174860760416256/  

AI governance may exist on paper. That does not mean it works when an autonomous agent is running in production. 

Louis Columbus, senior cybersecurity contributor at VentureBeat, joins Jo Peterson to talk about what happens when AI agents have identities, credentials, access and enough autonomy to act without waiting for human approval. 

They get into the gap between governance and enforcement, why companies may have far more agents running than they realize, and what CISOs need to know about agent identity, least privilege and credential risk. 

Louis also shares a case where an AI agent began modifying security policies to give itself more freedom. 

That changes the conversation. 

In this episode: 

• Why AI governance can become security theater 
• The identity problem created by autonomous agents 
• Why agent inventory has to come first 
• API keys, credentials and overprivileged access 
• What runtime enforcement actually means 
• Why ownership matters as much as policy 
• How enterprises can move toward “velocity with governance” 

About Louis Columbus 

Louis Columbus is a senior cybersecurity contributor at VentureBeat covering cybersecurity, enterprise AI, identity, zero trust and the security risks emerging as autonomous AI systems move into production. 

Read Louis’s work:
https://venturebeat.com/author/louis-columbus/ 

Additional Resources 

The credential that let OpenAI’s agents into Hugging Face exists in most enterprises right now
https://venturebeat.com/security/the-credential-that-let-openais-agents-into-hugging-face-exists-in-most-enterprises-right-now 

An AI now judges every move Rubrik’s agents make
https://venturebeat.com/security/an-ai-now-judges-every-move-rubriks-agents-make-its-ai-chief-said-at-vb-transform-2026-but-no-ones-measured-if-the-judge-is-right 

Four big enterprise lessons from Walmart’s AI security
https://venturebeat.com/ai/four-big-enterprise-lessons-from-walmart-ai-security-agentic-risks-identity-reboot-velocity-with-governance-and-ai-vs-ai-defense 

The Agent Security Gap
https://venturebeat.com/resources/the-agent-security-gap-54-of-enterprises-have-already-had-an-ai-agent-incident-and-most-still-let-agents-share-credentials 

Follow ClearTech Research 

Subscribe to ClearTech Loop on LinkedIn:
https://www.linkedin.com/newsletters/7346174860760416256/ 

Watch ClearTech Research on YouTube:
https://www.youtube.com/@ClearTechResearch 

Learn more:
https://cleartechresearch.com/ 

🎧 Listen: In Buzzsprout Player
▶ Watch on YouTube: https://www.youtube.com/@ClearTechResearch/videos
📰 Subscribe to the Newsletter:
https://www.linkedin.com/newsletters/7346174860760416256/  

Your AI agent has one job: accomplish the objective. 

It doesn’t care about your policies. It doesn’t worry about getting fired. And it may not stop when a human would. 

Joanna Wiggum calls it a narcissist. 

That sounds funny until you think about what happens when that “narcissist” has credentials, access to enterprise systems and permission to act autonomously. 

In this episode, Joanna joins Jo Peterson to talk about what AI agents are exposing inside enterprise security — weak credentials, unfinished zero-trust programs, governance committees with no real authority and security debt that AI can exploit at machine speed. 

Why Joanna says AI agents need “actual zero trust” 

Why some AI governance programs may be little more than security theater 

What happens when agents can create or delegate permissions 

Why old security failures become much more dangerous with autonomous AI 

What a CISO should do when the organization is already behind 

And when Jo asks Joanna what a CISO starting from zero should do in the next 30 days, Joanna doesn’t hesitate: 

“Outsource.” 

Listen to the full conversation to hear why. 

Short Summary 

AI agents don’t care about your policies. Joanna Wiggum joins Jo Peterson to explain why autonomous AI is exposing security debt, where zero trust breaks down and why some AI governance programs may be little more than theater. 

Links & Resources 

Countervail: https://countervailintelligence.com/ 

Joanna Wiggum — The Moral Imperative to Fight: https://countervailintelligence.com/2026/07/08/the-moral-imperative-to-fight/ 

OWASP — Agentic Security Initiative: https://genai.owasp.org/initiatives/agentic-security-initiative/ 

Related CTL: AI Agent Governance Starts With Visibility — Alvaro Gonzalez: https://cleartechresearch.com/ai-agent-governance-alvaro-gonzalez/ 

ClearTech Loop LinkedIn Newsletter: https://www.linkedin.com/newsletters/7346174860760416256/ 

ClearTech Research on YouTube: https://www.youtube.com/@ClearTechResearch 

🎧 Listen: In Buzzsprout Player
▶ Watch on YouTube: https://www.youtube.com/@ClearTechResearch/videos
📰 Subscribe to the Newsletter:
https://www.linkedin.com/newsletters/7346174860760416256/  

Guest 

Benny Czarny 
CEO, Founder & Chairman of the Board 
OPSWAT 

Host 

Jo Peterson 
CIO, Clarify360 
Chief Analyst, ClearTech Research 

AI agents have credentials. They access enterprise data. They make decisions. And increasingly, they can act without waiting for a human. 

So who actually owns the risk? 

Jo Peterson sits down with Benny Czarny, CEO, Founder & Chairman of the Board at OPSWAT, to talk about what AI agent governance needs to look like as autonomous AI moves deeper into the enterprise. 

They discuss why every AI agent needs its own identity and a human owner, how least privilege should apply to agents and sub-agents, and why understanding what data an agent can access may be just as important as securing the model itself. 

Benny also shares where CISOs should start if they don’t yet have an AI agent inventory or governance framework in place. 

In This Episode 

• Why every AI agent needs a human owner 
• Least privilege for AI agents and sub-agents 
• The growing risk around AI data access 
• Why AI governance needs real authority 
• The first step CISOs should take in the next 30 days 

Chapter Markers 

00:00 Meet Benny Czarny of OPSWAT
01:40 Protecting critical infrastructure
02:05 Least privilege for AI agents
03:58 Permanent credentials and autonomous agents
05:15 Does AI governance actually work?
07:23 Where CISOs should start
09:29 Protecting the AI data lake
12:34 What AI can learn from OT security
13:12 Cloud, on-prem and air-gapped AI
13:54 OPSWAT’s own AI journey 

Resource Links 

• OPSWAT https://www.opswat.com/ 
• OPSWAT Academy https://opswatacademy.com/ 
• Cybersecurity Upside Down — Benny Czarny https://www.amazon.com/dp/B0GH8SZXJ9 
• ClearTech Research https://cleartechresearch.com/ 
• ClearTech Loop Newsletter https://www.linkedin.com/newsletters/7346174860760416256/ 
• Full ClearTech Research Article https://cleartechresearch.com/ai-agent-governance-benny-czarny/ 

🎧 Listen: In Buzzsprout Player
▶ Watch on YouTube: https://www.youtube.com/@ClearTechResearch/videos
📰 Subscribe to the Newsletter:
https://www.linkedin.com/newsletters/7346174860760416256/  

AI agents can hold credentials, access sensitive data, make decisions and even create other identities. But many organizations still cannot answer a basic question: what is actually running in the environment? 

In this episode of ClearTech Loop, Jo Peterson sits down with Alvaro Gonzalez, SVP of Product and Go-to-Market at Assured Data Protection, to talk about what AI governance looks like when identity and access are changing at machine speed. 

Alvaro explains why organizations should start with three things: inventory, observability and remediation. They also discuss whether AI governance committees are actually governing or merely documenting, why CISOs should inventory agents before building more policy, and how Alvaro’s idea of “controlled aggression” can help enterprises experiment with AI without losing the ability to recover when something goes wrong. 

You cannot govern what you cannot see. 

Listen to Learn 

• Why AI agent governance should start with inventory 
• What least privilege looks like when identities can create other identities 
• Why observability matters alongside access control 
• Why remediation belongs in the AI identity conversation 
• Whether AI governance committees are really changing behavior 
• What Alvaro means by “librarians and warriors” 
• How “controlled aggression” can help organizations move faster without losing control 

Featured Soundbite 

“You cannot govern what you cannot see.” 

— Alvaro Gonzalez 

Featured Guest 

Alvaro Gonzalez 
SVP of Product and Go-to-Market 
Assured Data Protection 

Alvaro leads product, alliance, marketing and go-to-market functions at Assured Data Protection, with a focus on data protection, cyber resilience and the systems that support field and channel execution. 

Host 

Jo Peterson 
CIO, Clarify360 
Chief Analyst, ClearTech Research 

Episode Links 

Full episode webpage:
https://cleartechresearch.com/ai-agent-governance-alvaro-gonzalez/ 

Subscribe to ClearTech Loop:
https://www.linkedin.com/newsletters/7346174860760416256/ 

Watch on YouTube:
https://www.youtube.com/@ClearTechResearch 

Additional Resources 

Assured Data Protection: 5 Ways You Can Improve Your Cyber Recovery Plan
https://assured-dp.com/guides/5-ways-you-can-improve-your-cyber-recovery-plan-with-assured-data-protection/ 

NIST AI Risk Management Framework
https://www.nist.gov/itl/ai-risk-management-framework 

Model Context Protocol — Security Best Practices
https://modelcontextprotocol.io/specification/draft/basic/security_best_practices 

Previous ClearTech Loop: AI Agents Shouldn’t Be Trusted by Default with Elliott Mattice
https://cleartechresearch.com/ai-governance-trust-elliott-mattice/ 

🎧 Listen: In Buzzsprout Player
▶ Watch on YouTube: https://www.youtube.com/@ClearTechResearch/videos
📰 Subscribe to the Newsletter:
https://www.linkedin.com/newsletters/7346174860760416256/