Palo Alto Networks has introduced Unit 42 Frontier AI Defense, a new service designed to help organizations prepare for a new generation of threats powered by frontier AI, capable of significantly accelerating the detection of vulnerabilities, the generation of exploits and the execution of autonomous cyberattacks.
The launch comes at a key moment for the sector. Frontier AI models—that is, the latest and most advanced artificial intelligence systems, including Anthropic’s Mythos model as part of Project Glasswing and OpenAI’s most recent models within the Trusted Access for Cyber program—no longer act solely as programming assistants or productivity tools. Their reasoning, code analysis, and automation capabilities allow them to behave like broad-spectrum security researchers, facilitating both the prevention and sophistication of cyberattacks.
As these models become more accessible, attackers will be able to automate critical phases of the attack and cyberattack cycle, such as reconnaissance, vulnerability identification, lateral movement, generating custom exploits, or prioritizing data for exfiltration. This scenario is especially relevant in environments with a high dependence on open source software, complex supply chains, services exposed to the internet and hybrid or multicloud architectures, where cyberattacks can escalate more quickly.
According to Unit 42, the risk is not limited to the discovery of zero-day vulnerabilities. An acceleration of N-day threats is also expected, those that exploit vulnerabilities already known but not yet corrected by organizations, increasing the volume of cyberattacks. In this context, the traditional patching window can go from days to hours, increasing the pressure on security, development and infrastructure teams to effectively mitigate cyberattacks.
Evaluate, prioritize and transform
To help companies respond to this new cyberattack scenario, Unit 42 Frontier AI Defense combines the expertise of Unit 42’s specialized consultants with the latest AI models, threat intelligence and telemetry from Palo Alto Networks. The service is divided into three main components:
Frontier AI Exposure Analysis: To identify and validate exposures most likely to be chained into real cyberattacks before they are exploited by adversaries. To do this, it combines advanced AI models, Unit 42’s offensive experience, threat intelligence and security telemetry, offering a prioritized view of vulnerabilities, configuration errors, posture gaps and attack paths in infrastructure, applications, code, identity and cloud.
Autonomous Security Blueprint: This component evaluates the organization’s current capabilities and defines the changes necessary to move towards a prepared defense against machine-speed cyberattacks. The analysis covers attack surface, identity, software supply chain, zero trust, real-time detection and incident response, generating a prioritized roadmap to reduce exposure, reinforce containment and modernize security operations against cyberattacks.
Unit 42 Frontier AI Defense combines the expertise of Unit 42’s specialist consultants with the latest AI models
Agentic Defense Transformation: The third phase focuses on accompanying the implementation of the architectural, operational and control changes necessary to defend against AI-driven threats, including new types of cyberattacks. This includes modernizing exposure management, strengthening the software supply chain, advancing zero trust architectures, and developing response capabilities capable of keeping pace with autonomous cyberattacks.
With Unit 42 Frontier AI Defense and the creation of the Frontier AI Alliance, together with global transformation leaders such as Accenture, Deloitte, IBM, NTT DATA and PwC, Palo Alto Networks reinforces its commitment to helping organizations anticipate this new era of cybersecurity, in which human-speed security will no longer be enough and resilience against AI-powered cyberattacks will become a strategic priority.
