This Week's Top Five Stories in Cyber

Ivanti's Chris Goettl: Preparing for the AI Patch Apocalypse

Among the industries that AI is reshaping, cybersecurity is one that is shifting at an unprecedented speed – transforming how vulnerabilities are discovered, disclosed, exploited and fixed across enterprise environments.

As frontier AI models like Mythos and GPT 5.4-cyber accelerate vulnerability research, security teams are facing mounting pressure to manage a surge in patches, rising cyber risk and shrinking response windows.

This growing challenge is exposing the limitations of traditional patch management and forcing organisations to adopt more agile, risk-based security strategies. 

In this Q&A with Cyber Magazine, Chris Goettl, VP of Product Management at Ivanti, explores how AI is changing the vulnerability lifecycle, why out-of-band patching is becoming more disruptive and how organisations can strengthen cyber resilience through continuous exposure management, automated remediation and smarter vulnerability prioritisation.

Moody's Warns that AI Cyber Arms Race Raises Risks for Banks

Frontier AI has changed the bug hunting game forever – unleashing a storm of vulnerabilities faster than organisations can patch them – forcing businesses into a new era of cyber resilience where defence speed and system architecture matter more than ever.

Among the industries that this shift is sending off to the cybersecurity battlefield, financial institutions are in the frontline dealing with the fallout. 

The panic started after the release of highly capable models such as Anthropic’s Claude Mythos and OpenAI’s GPT-5.4-Cyber that are capable of autonomously identifying thousands of previously unknown software vulnerabilities across major operating systems and browsers.

Large US banks were among the first to gain controlled access to Mythos through Project Glasswing, an initiative designed to stress test cyber defences against next-generation AI threats. 

A recent Moody’s industry report titled “Arms Race: Deep defenses will help banks navigate cyber threats from new AI models” dives into the cyber risk landscape for financial institutions.  

Marcus Lauren on battling Deepfakes with Active Thermal tech

It is no secret that our mobile phones and laptops are home to sensitive data, all the way from corporate budgets and colleague details, to personal banking information.

The responsibility to protect this data rests with biometric technology, which is relied upon constantly, dozens of times each day.

As Gen AI and deepfakes become more capable than ever before, the digital security perimeter has altered and modern cybersecurity architecture must adapt promptly to counter these threats. Zero Trust frameworks address this requirement.

John Kindervag, Chief Evangelist at Illumio –who is widely acknowledged as the original creator of the Zero Trust framework – states the core strategy must be to "never trust, always verify".

In this Q&A, Marcus Lauren, Chief Product Officer at NEXT Biometrics, discusses how the company is redefining identity and access management (IAM), including the critical role of physical biometrics in an AI-driven world. 

How Mistral AI Drives Sovereign AI Adoption in Manufacturing

AI is accelerating across industry at pace, but for cybersecurity leaders in manufacturing and critical infrastructure it introduces a familiar tension.

General-purpose models are rarely tuned for industrial environments yet exposing proprietary engineering data to train specialised systems can significantly increase cyber risk and intellectual property leakage.

A wave of partnership announcements with European AI firm Mistral AI highlights how manufacturers are navigating that balance.

Major industrial players including Airbus and BMW Group confirmed collaborations on 28 May, following earlier agreements with Stellantis and ASML.

From a cyber perspective these deals raise an important question: how do organisations harness high-performance AI without expanding their attack surface or losing control over sensitive operational data?

How Cisco Protects AI Agents From the World of Cyber Threats

As AI agents become a more central part of the enterprise workforce, a new system of command and control is necessary. 

Appealing to this market-wide sentiment, tech giant Cisco has launched Cloud Control, a new unified platform designed to help organisations manage and protect critical IT infrastructure in an increasingly AI-driven world. 

Unveiled at Cisco Live US 2026 in Las Vegas, the platform forms the foundation of Cisco's AgenticOps vision, enabling human operators and AI agents to work together from a single environment. 

Cloud Control provides a consolidated view of networking, security, compute, observability and collaboration, all through a single environment. 

Built around a shared data layer, the platform allows people and AI agents to operate with the same information and context while ensuring that humans remain in control of decision-making.