Open Source in the AI Era
Modern software is built predominantly using open source components. As development teams introduce agentic workflows into their build processes, AI agents are choosing the open source dependencies. But those choices can introduce dangerous software. In this video, Sonatype CTO and co-founder Brian Fox explains why dependency decisions matter and how teams can guide AI agents toward better choices, verify what they select, and stop risky components before they cause harm.
Related Resources
The Accelerated AI Threat
Frequently Asked Questions
What are open source dependencies?
Open source dependencies are libraries, frameworks, and other reusable components that a software project relies on. They help developers build on existing code instead of creating every feature from scratch. A dependency can also bring in its own dependencies, creating a larger chain of components.
Why do dependency choices matter?
Dependencies affect an application’s security, quality, maintenance, and licensing. A component may have known vulnerabilities, be outdated or unmaintained, or carry a license that could jeopardize your application and your organization.Because dependencies can bring in other components, one choice can affect more of the application than expected.
How do AI agents make dependency choices?
AI agents use a prompt, available tools, and the surrounding code or project context to recommend or add dependencies. The choice may depend on patterns in the agent’s training data or information it can retrieve. If the agent lacks current, reliable component details or project-specific guidance, it may suggest a component that is unsafe, poor quality, or nonexistent.