The CISO Perspective: Cybersecurity, AI and Emerging Tech

AI agents now reach your email, files and business apps through the Model Context Protocol (MCP), and Backslash Security found that nearly a third of the most popular MCP servers carried at least one security risk finding. This episode explains how MCP works and walks through the 34 threats the Coalition for Secure AI (CoSAI) lists in its MCP security framework.Key takeaways:• What MCP is: a standard way for AI agents to discover and use tools, either natively or through a separate MCP server• Why connecting an agent to untrusted content opens a new pathway into the corporate environment• Conventional threats (19): weak authentication, excessive privileges, insecure network configuration, poor credential management• MCP-contextualized threats (8): prompt injection, confused deputy, denial of wallet, man-in-the-middle, consent fatigue• MCP-specific threats (7): identity spoofing, tool poisoning, full schema poisoning, resource content poisoning, typosquatting, shadow MCP servers, over-reliance on the LLM• You don't need to reinvent security for AI agents; extend least privilege, strong authentication, segmentation and vulnerability management, and avoid "slop security"Chapters:0:00 Intro: why MCP security matters1:04 What is MCP?2:17 The CoSAI framework3:46 Conventional threats5:34 MCP-contextualized threats7:06 MCP-specific threats8:51 Wrapping up: trust and "slop security"9:17 ClosingSources:• Backslash Security, MCP server security research (Sep 2026): https://www.backslash.security/blog/mcp-server-security-risks-research• CoSAI, Model Context Protocol Security v2.0 (Sep 2026): https://www.coalitionforsecureai.org/wp-content/uploads/2026/09/model-context-protocol-security-v2.0.pdf• Snyk, malicious "postmark-mcp" npm package (Sep 2025): https://snyk.io/blog/malicious-mcp-server-on-npm-postmark-mcp-harvests-emails/Watch the full video: https://youtu.be/iD6oqZV8AlEMore from The CISO Perspective: https://cisoperspective.com

Verizon's 2026 DBIR found the human element in 62% of breaches, and increasingly the browser is where users get tricked into running the attack themselves. This episode compares remote browser isolation (RBI) and the secure enterprise browser: how each works, what problem each solves, and where they fit in SASE.

Key takeaways:
• How RBI runs risky pages in a disposable cloud browser (pixel streaming vs DOM reconstruction)
• Why RBI is applied selectively through a secure web gateway or SASE platform
• How a secure enterprise browser puts policy inside the session: copy/paste, uploads, downloads, credentials
• Rule of thumb: isolate when the page is the threat; secure the browser when the work is trusted but the user, device or data needs control

Chapters:
0:00 Intro: the human element and the browser
1:07 The two options: RBI vs secure enterprise browsing
1:54 How remote browser isolation (RBI) works
2:33 Pixel streaming vs DOM reconstruction
3:10 Why RBI is applied selectively
4:11 How secure enterprise browsing works
5:03 Policy controls inside the browser session
5:52 So which one do you need?
6:25 Where they fit: SASE and Gartner's outlook
6:52 The rule of thumb

Watch the full video: https://youtu.be/ttmgM_AbkgU
More from The CISO Perspective: https://cisoperspective.com

Security teams can find vulnerabilities faster than ever. The hard part is deciding which exposures actually matter. This episode breaks down Continuous Threat Exposure Management (CTEM), Gartner's five-stage program for focusing on the exposures that create real business risk.Key takeaways:• CTEM is a program, not a product you buy• The five stages: scoping, discovery, prioritization, validation and mobilization• Prioritizing with KEV and EPSS instead of trying to fix everything• Why the cycle has to stay continuousChapters:0:00 The vulnerability volume problem0:25 What is CTEM?1:54 The five stages at a glance2:10 Stage 1 - Scoping2:57 Stage 2 - Discovery3:54 Stage 3 - Prioritization5:00 Stage 4 - Validation6:01 Stage 5 - Mobilization6:32 The goal of CTEM7:08 Why the cycle stays continuousWatch the full video: https://youtu.be/Mu0PcfaHLEoMore from The CISO Perspective: https://cisoperspective.com

Extended Detection and Response (XDR) is a cybersecurity tool that integrates with multiple products to detect and respond to threats. While EDR (Endpoint Detection and Response) focuses on the endpoint, XDR uses A.I. to ingests information from multiple sources and detect patterns across different device types.
You’ll learn:
What EDR does: how it monitors endpoint activity and helps isolate threats in real time 
How MDR elevates EDR by providing 24/7 managed monitoring, threat hunting, and guided response from security experts 
How XDR takes it further—correlating data across endpoints, network, cloud, identities, and more to give a unified view and faster response   

A clear breakdown of each detection layer
A comparison of when to use which solution (or combination)
Practical tips for choosing based on your organization’s expertise and infrastructure

Whether you’re defining your security strategy or just curious about detection platforms, this episode gives you:Topics covered: EDR functionality, MDR benefits, XDR scope, vendor considerations, and strategic fit for CISOs and security practitioners. Tune in to sharpen your detection stack—without the jargon.

How do you turn Zero Trust theory into real-world implementation?

In this episode of The CISO Perspective, we break down the actual technologies and solutions that power modern Zero Trust Architecture (ZTA). You’ll hear how tools like identity providers, next-gen firewalls, endpoint protection, and segmentation gateways all work together to meet the core principles of NIST SP 800-207.

We’ll explore how each technology brings unique strengths to the framework — and more importantly, how they integrate to build a practical, enforceable Zero Trust environment.

👉 This is Part 2 of our Zero Trust series. If you missed Part 1, be sure to check it out for an overview of the key components, design principles, and architectural models that define Zero Trust.