Goal Summary:

Deliver a Tech Preview of AI workload identification and AI BOM ingestion in RHACS, implementing one of the architectural approaches discussed in the Discovery phase. This version provides a functional demonstration to customers, enabling feedback and validation, but is not yet production-ready. The Tech Preview includes identifying AI workloads, ingesting AI BOMs, and associating metadata with workloads in RHACS.

Goals and expected user outcomes:

•  
• Implement a functional Tech Preview of AI workload identification and AI BOM ingestion using one of the options evaluated in Discovery (classification as existing workloads, new AI workload abstraction, or hybrid).

• Enable end users to explore, test, and provide feedback on AI workload identification and BOM ingestion capabilities.

• Demonstrate how AI BOM metadata can be associated with workloads in RHACS.

• Validate architectural feasibility, integration patterns, and data handling approaches in a real cluster environment.

End user outcome:

• Users can see which workloads in their clusters contain AI/ML models.

• Users have a centralized view of AI models and associated metadata through RHACS.

• Users can attach externally generated AI BOMs to workloads or images and explore associated metadata.

• Users can provide feedback on the Tech Preview to influence future GA implementation.

One of the options from Discovery will be implemented for Tech Preview:

• • Classification as existing Kubernetes objects with AI metadata

• • New AI workload abstraction with AI metadata

• • Hybrid approach with metadata attached to existing workloads

The specific approach selected will be documented and justified based on Discovery findings.

Acceptance Criteria:

• Users can explore AI workload identification and BOM metadata in RHACS.

• Working examples  showcasing AI workload identification and AI BOM ingestion.

Success Criteria or KPIs measured:

• Completion and deployment of a functional Tech Preview demonstrating AI workload identification and BOM ingestion.

• Users are able to interact with AI workload identification and view associated AI BOM metadata.

Use Cases (Optional):

• Security Engineer: Can explore AI workloads and associated BOM metadata to assess potential governance needs.

• Cluster Administrator: Can inventory AI workloads and associated metadata to test management and compliance workflows.

• DevSecOps Team Member: Can attach sample AI BOMs to workloads or images and evaluate integration with CI/CD pipelines.

Out of Scope (Optional):

High-level list of items that are out of scope. Initial completion during Refinement status.

• Delivery of production-ready AI workload discovery or AI BOM ingestion.

• Vulnerability scanning or CVE correlation for AI models.

• Behavioral, safety, bias, fairness, or hallucination evaluation of AI models.

• Native execution, deserialization, or analysis of AI models within RHACS.

• Integration with specific third-party AI artifact scanners for production use.

• Performance, scalability, or reliability guarantees beyond demonstration.

• Future GA enhancements (policy enforcement, broader scanner integration) — to be addressed in subsequent feature.