Palo Alto Networks has conducted early testing of the latest Frontier AI models, including Anthropic’s Mythos model as part of Project Glasswing and OpenAI’s most recent models within the Trusted Access for Cyber program.
The company has confirmed that the generational improvement in the programming capabilities of frontier AI directly translates into significant progress in the detection of vulnerabilities and the generation of exploits.
The capabilities of frontier AI, while protected by safeguards, will not remain contained. Attackers will find loopholes in those safeguards and use cutting-edge AI to discover zero-day vulnerabilities at scale, generate exploits in near real-time, and develop autonomous attack agents unlike anything the industry has ever seen before.
“Within six months, advanced frontier AI models with deep cybersecurity capabilities will be commonplace, and organizations that have not implemented adequate safeguards will face an entirely new class of risks across their enterprise and critical infrastructure,” say Palo Alto Networks experts.
Frontier AI and its impact on the cybersecurity landscape
Hundreds of Palo Alto Networks’ top security engineers have evaluated these frontier AI capabilities and developed best practices for using them effectively, finding that:
- Frontier AI is exceptionally effective at identifying vulnerabilities in code. In less than three weeks, he accomplished a full year’s worth of penetration testing work.
- Even more impressive, frontier AI excels at chaining vulnerabilities, combining multiple lower-severity flaws into critical-level exploit paths.
- Edge AI can analyze the entire exposure surface of applications, including SaaS platforms and internet-exposed services, identifying logic-based vulnerabilities that traditional tools miss.
Key areas where frontier AI will have a significant impact on cybersecurity
Frontier AI models will dramatically accelerate the pace at which vulnerabilities are discovered. This will be especially critical in open source, and the avalanche of patches that will follow will create new risks.
Rise of inside-out attacks: Recent supply chain attacks on tools like LiteLLM and Trivy demonstrate a growing pattern in which attacks place adversaries inside an organization’s infrastructure, bypassing multiple traditional attack phases and reducing prevention opportunities for defenders. Frontier AI will amplify this risk.
The key change with frontier AI models is the move from AI-assisted attacks to AI-powered attacks. Attackers will build autonomous attack agents that will dramatically compress attack cycle times. What once required days or weeks of specialized manual effort will soon be executed in minutes thanks to cutting-edge AI capabilities.
Guide for defenders
Organizations that are “mostly protected” are “effectively unprotected,” say Palo Alto Networks experts, offering recommendations on the framework needed to defend against frontier AI-driven threats:
- Every organization should use the most advanced frontier AI models to assess their entire code and application environment, and build a comprehensive asset and exposure inventory, focusing on priorities such as identifying vulnerabilities with AI, assessing exposure with full context, auditing the open source supply chain, and mapping current sensor coverage.
- Remediation and exposure reduction are basic requirements. What was difficult in the past due to friction between teams in quickly detecting and fixing vulnerabilities must now be accelerated with the executive committee’s attention to these new frontier AI models. However, it is necessary to go further and comprehensively implement world-class attack prevention capabilities.
The key change with frontier AI models is the move from AI-assisted attacks to AI-powered attacks
- With the accelerated reduction of attack cycles driven by cutting-edge AI, the traditional approach to security operations is no longer valid. Siled tools that analyze data in silos, combined with manual processes, must be replaced by AI and end-to-end automation.
Palo Alto Networks is driving the industry standard to address these emerging risks with its Unit 42 Frontier AI Defense, focused specifically on threats arising from frontier AI. The company is also advancing an alliance of global transformation leaders, starting with Accenture, Deloitte, IBM, NTT DATA and PwC, which it will continue to expand to ensure that all companies have a fast track to resilience in frontier AI.
