Daily Cyber & AI Briefing with Michael Housch. This episode was published automatically and includes the assembled audio plus full transcript.
Today’s cyber and AI risk landscape is more dynamic and complex than ever, with both technical exploits and governance challenges converging to create a high-pressure environment for security leaders and risk executives. We’re seeing a surge in high-profile vulnerabilities, active exploitation by threat actors, and a rapidly evolving regulatory climate driven by the widespread adoption of artificial intelligence. Let’s break down what’s happening, why it matters, and what organizations need to do to stay ahead.
First, let’s talk about the wave of technical exploits targeting enterprise infrastructure. Ransomware groups are actively exploiting a vulnerability in Palo Alto Networks’ GlobalProtect VPN. This isn’t just another bug—it’s being used to bypass authentication and maintain persistent, stealthy access to enterprise environments. The implications are serious: attackers can move laterally, exfiltrate sensitive data, and deploy ransomware, all while evading detection for longer periods. For organizations relying on GlobalProtect, the message is clear: prioritize patching, monitor VPN activity for anomalies, and rigorously review remote access controls. This incident underscores a broader trend—remote access infrastructure remains a prime target for sophisticated threat actors, and the window between vulnerability disclosure and active exploitation is shrinking.
The situation with SonicWall’s SMA 1000 series is another stark example. Even after a patch was released for a critical vulnerability, there have been at least 135 documented exploitation attempts. Attackers are quick to weaponize newly disclosed flaws, often beating organizations to the punch on patching. This highlights the need for not just rapid patch deployment, but also vigilant post-patch monitoring. It’s not enough to apply a fix and move on—security teams need to verify that patches are actually in place, monitor for indicators of compromise, and be prepared for attackers who may have gained a foothold before remediation.
Web-facing applications are also under siege. The U.S. Cybersecurity and Infrastructure Security Agency, or CISA, has added an actively exploited command injection vulnerability in Apache Struts to its Known Exploited Vulnerabilities catalog. This is a high-severity issue that enables remote code execution, posing a significant threat to any organization running Struts-powered applications. CISA’s inclusion of this flaw in its KEV catalog isn’t just a bureaucratic step—it’s a call to immediate action for federal agencies and a strong recommendation for all enterprises. If you’re running Struts, urgent remediation is non-negotiable.
Open-source platforms are not immune, either. Discourse, a widely used discussion platform, has been found to contain multiple vulnerabilities that allow attackers to execute arbitrary JavaScript and read sensitive server files. These kinds of flaws can be exploited to compromise user data and escalate privileges, making them attractive targets for attackers. Organizations using Discourse should apply patches as soon as they’re available and take this as a reminder to regularly review their web application security posture. Open-source tools are valuable, but they require the same level of scrutiny and maintenance as any commercial solution.
As we move from technical exploits to the evolving world of AI, the landscape becomes even more nuanced. The adoption of AI across business functions is accelerating, but so are the risks and regulatory demands. Rubrik’s expansion of Project Hourglass is a timely example. They’re now focusing on testing AI-generated code for cybersecurity vulnerabilities. As more organizations leverage AI to write and review code, the risk of introducing exploitable flaws increases. Automated code review tools that specifically address AI-generated code are becoming essential. Security leaders should integrate these tools into their secure development lifecycle and update practices to account for the unique risks posed by AI-assisted development.
The regulatory environment is also shifting rapidly. Managed service providers, or MSPs, are facing new certification demands as their use of AI grows. Customers and regulators alike are demanding proof of AI governance, robust security controls, and responsible use policies. This trend is not limited to MSPs—any organization deploying AI should expect similar scrutiny. CISOs need to stay on top of evolving standards and ensure their teams are prepared for audits and compliance reviews. The days of deploying AI in a regulatory vacuum are over.
Nowhere is this more evident than in the European Union, where the EU AI Act is poised to reshape how organizations govern and deploy AI. The Act introduces significant requirements around risk management, transparency, and accountability. Non-compliance isn’t just a matter of regulatory slap on the wrist—it can result in substantial fines and reputational damage. Corporate leaders should be proactively assessing their AI systems for compliance, updating documentation, and implementing robust governance frameworks. Early preparation is essential, not just for legal compliance, but for maintaining stakeholder trust.
The integration of AI into identity governance is another area to watch closely. Anugal’s introduction of agentic AI capabilities into enterprise identity governance brings both promise and peril. On one hand, more autonomous and adaptive access decisions can drive efficiency and agility. On the other, these systems raise new questions about explainability, oversight, and the potential for abuse. Security teams must ensure that AI-driven identity systems are auditable, transparent, and aligned with organizational policies. The balance between automation and control is delicate, and getting it wrong can expose organizations to significant risk.
This brings us to a broader and increasingly urgent issue: the rise of AI agents as a new class of insider threat. Traditionally, insider threat programs have focused on human actors—employees, contractors, and partners with access to sensitive data and systems. But as AI agents proliferate within organizations, they too can access sensitive data, make decisions, and potentially be manipulated or subverted. CISOs should update their insider threat programs to account for these non-human actors, ensuring that monitoring and controls are in place for AI-driven processes. The line between human and machine insiders is blurring, and security strategies need to evolve accordingly.
Recent events have also highlighted the importance of trust and vetting in third-party relationships. The arrest of a ransomware response expert in connection with the ShinyHunters FBI hack investigation is a sobering reminder that even trusted incident response providers can become sources of risk. Security leaders should review their vendor management and incident response protocols to ensure integrity and compliance. This includes not just technical vetting, but also ongoing monitoring and clear contractual requirements around security and ethics.
Amidst all these challenges, collaboration is emerging as a critical enabler of cyber resilience. LevelBlue is emphasizing the importance of collaboration across teams, partners, and even sectors to build resilience and support business growth. As threats become more complex and interconnected, siloed approaches are increasingly ineffective. CISOs should foster cross-functional partnerships and encourage information sharing, both internally and with external partners. The ability to respond quickly and effectively to incidents often depends on the strength of these collaborative relationships.
Zooming out even further, a comprehensive risk assessment has identified AI misuse as one of 17 existential risks facing humanity, alongside climate change and other systemic threats. For risk executives, this underscores the need to integrate AI risk into broader enterprise risk management frameworks and scenario planning. It’s not just about compliance or technical controls—it’s about understanding and preparing for the full spectrum of risks that AI can introduce, from operational disruptions to reputational crises and even societal impacts.
So, what are the key imperatives for risk executives and security leaders today?
First, maintain relentless vigilance on patching and vulnerability management. The rapid exploitation of newly disclosed vulnerabilities means organizations must have processes in place to identify, prioritize, and remediate critical flaws—especially in infrastructure like VPNs, firewalls, and web applications. But patching alone isn’t enough. Continuous monitoring for signs of exploitation, even after patches are applied, is essential to catch attackers who may have already gained access.
Second, prepare for the evolving landscape of AI regulations and certifications. This means assessing your AI systems for compliance, updating governance frameworks, and ensuring that your teams are ready for audits and certification reviews. The regulatory environment is only going to become more demanding, and early preparation will pay dividends in both compliance and operational resilience.
Third, expand your insider threat programs to include AI agents and non-human actors. As AI becomes more deeply embedded in business processes, it’s critical to recognize that these systems can introduce new vectors for insider risk. Monitoring, controls, and oversight mechanisms need to evolve to keep pace with this shift.
Fourth, foster a culture of cyber resilience that spans technology, process, and people. This includes upskilling your workforce, encouraging cross-