New Chainguard Academy course: Securing the AI/ML Supply Chain

With the artificial intelligence (AI) revolution underway, we now face a similar turning point we saw with the rise of open source software adoption and development. If the industry doesn’t proactively bake security into popular tooling and practices for AI and machine learning (ML) development, we will build a hole-ridden foundation that will have impacts for decades to come. Learn more about this turning point in our CEO Dan Lorenc’s recent AI_Dev keynote address.

‍

To help the industry learn and understand best practices for securing AI/ML, Chainguard is launching a hands-on, self-paced course — Securing the AI/ML Supply Chain — designed for AI/ML developers, security professionals, and other stakeholders.

‍

With AI technologies bringing autonomous decision making into critical areas like transportation, defense, and medicine (to name only a few), the stakes are higher than they’ve ever been before and a safer foundation for this wave of development is critical.

‍

What is the AI/ML supply chain?

‍

Like traditional software, AI/ML systems are also the product of a long, complex supply chain of data, dependencies, infrastructure, distribution systems, and development tools. By approaching AI/ML security from a supply chain perspective, we are able to take a broad and holistic view of the key ways AI/ML systems can be compromised, including attacks via a system’s model, the training data, or its software infrastructure.

‍

This course will teach you the fundamentals of supply chain security concepts and practices for AI/ML development, and will include hands-on demonstrations of key tools for mitigating risk. Part I of this course covers the AI/ML supply chain's components, the different profiles and motivations of cyberattackers, the range of cyberattacks targeting AI/ML supply chains, and evolving international regulations, standards, and frameworks relevant to the field.

‍

Part II provides a variety of hands-on demonstrations of tools and practices for software supply chain security that can be applied to AI/ML development. These lessons include exercises on scanning AI/ML images and other relevant software for vulnerabilities as well as creating and consuming SBOMs, digital signatures, and provenance documents. You will also learn how to use our minimal, secure Chainguard AI Images (such as PyTorch) to drastically reduce your AI/ML system’s vulnerabilities and size.

Participants who finish the course will be awarded Chainguard’s Securing the AI/ML Supply Chain badge on Credly and a certificate, both of which can be shared on LinkedIn and other social platforms. More importantly, however, participants will come away with the skills and resources to improve the software supply chain security of their AI/ML systems and share their knowledge with their organization. As the industry embarks on this next stage of AI/ML innovation, let’s make sure that this time we get security right.

‍

Getting started with Chainguard AI Images

‍

Our Securing the AI/ML Supply Chain course is also a great precursor to get started with our Chainguard AI Images, a growing suite of CPU and GPU-enabled container images, including PyTorch, Conda, Kafka, that are hardened, minimal, and optimized for efficient software development. These Images are available to use starting today!

‍

Register for Chainguard’s Securing the AI/ML Supply Chain course today. Not ready to sign up? Check out our AI Guide — Is Your AI Trustworthy? Unmasking the hidden dangers of the AI/ML supply chain — to get you started on your journey to more secure AI/ML development.