Open Secure AI Alliance Launches asago Project for Automated AI Governance

The Open Secure AI Alliance has launched an open-source project called asago that aims to automate the link between AI governance policy and production-ready deployment code. Red Hat, NVIDIA, and IBM are among the founding members of this project.

Asago is a four-stage workflow that first maps organizational governance policies against established frameworks like NIST AI RMF, OWASP LLM Top 10, and the EU AI Act. It then generates scenarios tailored to specific use cases, probes for harmful behaviors, and recommends guardrails based on testing results.

The project's outputs are meant to be infrastructure-agnostic, with declarative configurations for Kubernetes, Terraform, and Ansible. This allows safety postures set in one cloud to not need re-engineering in another.

Red Hat aims to reduce deployment timelines from months to days using asago. The project's founders believe that AI safety is an engineering problem as much as a policy problem.