Application Security
,
Artificial Intelligence & Machine Learning
,
Events
Secure Code Warrior’s Pieter Danhieux on Managing AI-Driven Development Risks
Software development is moving from human-led to agent-led at a pace that security organizations are not built to absorb. Developers frequently switch platforms, making it difficult for security leaders to track usage or enforce policies, said Pieter Danhieux, co-founder and CEO of Secure Code Warrior.
See Also: How Cyber Deterioration Raises Enterprise Risk
Danhieux said visibility into approved artificial intelligence models and the model context protocols agents use to connect to APIs is essential. Without it, organizations risk exposing sensitive data or introducing vulnerable code into production. Code quality itself remains difficult to assess, because output depends heavily on how developers instruct the tools.
“If you do that in a proper way, then the agent or the LLM is going to produce code that is vulnerability-free. But do that the wrong way, or speak to it in language where you don’t give the right instructions, it might give much worse results much faster,” he said.
In this video interview with Information Security Media Group at RSAC Conference 2026, Danhieux also discussed:
- How developer AI adoption follows a maturity curve, and why CISOs need to know where their teams sit on it;
- How secure-by-design principles improve AI-generated code outcomes;
- Why the right governance posture is controlled adoption, not fast adoption.
At Secure Code Warrior, Danhieux leads efforts to reshape how the world approaches security in an AI-driven software development environment. He is also a principal instructor at the SANS Institute, where he teaches military, government and private firms offensive techniques on targeting and assessing organizations, systems and individuals for security weaknesses.
Click Here For The Original Source.
