LLMs as a New Attack Surface: what does it mean for AI governance?

Large Language Models (LLMs) are transforming industries, but their unique risks demand a new approach to security and governance. A groundbreaking paper co-authored by Anove Co-Founder Prof. dr. Yuri Bobbert and ethical hacker Kevin Zwaan from Q-Cyber exposes how traditional security controls fall short when AI behavior can be steered through plain everyday language.

A recent demonstration showed how an LLM could be "radicalized" over eight hours, bypassing safety guardrails to generate malware at scale. This wasn't a highly technical code-written software exploit; it was achieved through manipulation and persuasion, taking advantage of the model’s contextual learning to make it unlearn its security protocols, revealing a critical gap in AI security.

The paper highlights that AI's attack surface is broader than code. It includes the model, prompts, user interfaces, policies, and even the organizational context. When LLMs are integrated into workflows with access to tools, APIs, and sensitive data, the risks multiply, ranging from generating malicious content to enabling large-scale cyberattacks. AI systems are dynamic, made up of interconnected components that evolve rapidly. As a result, traditional governance can’t keep up. Static checklists and one-time audits aren’t enough (if they ever were). AI management must be continuous, automated, and evidence-based.

Read the full paper here to dive deeper into how to manage the risks of adopting AI in your business and how Anove’s AI Management System and Q-Cyber can enable responsible AI at scale. As the authors summarize, "the faster innovation moves, the more governance must become automated and operational—otherwise it becomes theater".