Beyond the Model Card: Engineering Trust into AI with AIBOM and CycloneDX

Join us for a meetup dedicated to Engineering Trust into AI Systems. Matt Rutkowski will present his talk "Beyond the Model Card".

Agenda:
First ~10 minutes - welcome, orientation, introductions
Next 30-40 minutes - Matt's talk "Beyond the Model Card: Engineering Trust into AI Systems using AI/ML-BOM and CycloneDX"
Last 10-20 minutes - Q&A, open discussion

Talk abstract:
Model cards tell a story about an AI model, but trust is relative. It's judged against the system or application a model is embedded in, and the use case it's asked to serve. As the EU AI Act and Cyber Resilience Act push general-purpose AI providers from voluntary disclosure toward mandatory, auditable technical documentation, the industry needs a structured, machine-readable answer, and not another PDF.

This talk draws on my work ensuring IBM's Granite models run across open and partner frameworks and apps — including local, agentic deployments — and on chairing the OWASP CycloneDX AI/ML Working Group, to show how CycloneDX AI/ML-BOMs can turn the concept of "model cards" into verifiable, regulation-mapped data. The goal: assess risk, confirm that the correct model — benchmarked and deployed for the task it's designed for — is actually in use, and verify regulatory and corporate policy compliance.

About Matt:
Matt Rutkowski is a Senior Technical Staff Member (STSM) with IBM Research, where he focuses on the Granite Partner Ecosystem, leading the development of and driving support for IBM's Granite models across both open source and partner frameworks and applications.

Representing IBM in open source and open standards communities has been a core part of his role for many years, and today he's a recognized leader in AI and machine learning security. Matt is a core maintainer of OWASP CycloneDX, a flagship OWASP project, where he leads the AI/ML Working Group and serves as lead author of the AI/ML-BOM Authoritative Guide, work that is defining how organizations achieve transparency and security across the AI/ML supply chain.

Event sponsor:
ReARM by Reliza - https://rearmhq.com