Kossi Messan
Director, AI Reliability Institute

Kossi Messan

"Defining the emerging role of the 'AI Reliability Professional' distinct from traditional safety engineering."

Expertise
AI Reliability Engineering
Role
Founder & Director
Focus
ISO/IEC 42001
Founded
2025

Current Contribution

Executive Director

Leading the Institute's mission to operationalize reliability for AI, Autonomous, and Agentic (AAA) systems.

Interim Chief Scientist

Overseeing the technical validation of AIR™-ES frameworks and reliability methodologies.

Chair, AIR™-CCWG (Standards)

AI Reliability Enforcement Standards (AIR™-ES) Development and Harmonization. Coordinating AIRI's contributions to international standards bodies and harmonizing AIR™-ES with global frameworks like ISO, NIST, and CEN.

Biography

Kossi Messan is the Founder and Director of the AI Reliability Institute (AIRI), a research organization dedicated to operationalizing reliability for AI, Autonomous, and Agentic (AAA) systems. Established in 2025, his work defines the emerging role of the "AI Reliability Professional," a discipline distinct from traditional safety engineering.

He authors foundational research designed to standardize how to test, measure, and enforce the reliability of AAA systems under frameworks like ISO/IEC 42001. His work focuses on bridging the gap between high-level regulatory requirements and binary engineering constraints.

ORCID iD icon https://orcid.org/0009-0002-1254-8586

Strategic Focus Areas

  • ISO/IEC JTC 1/SC 42 Strategy: Aligning AIRI frameworks with international AI standards.
  • NIST & RMF Alignment: Ensuring compatibility with the Risk Management Framework.
  • AIR™-ES Version Control: Managing the evolution of Enforcement Standards.

Key Deliverables

  • Shadow Draft Proposals

    Technical proposals for standards bodies.

  • AI Reliability Lexicon

    Defining the standardized vocabulary for Agentic Control.

  • Standards Consultation Responses

    Official responses to public consultation periods for EU AI Act & NIST frameworks.