Description

A recent whitepaper by the founder of a fledgling think tank identifies three common failure modes in agentic AI governance, chief among them the idea that AI models can be trusted to govern themselves. That founder, and the six companies he's founded, propose a new way to more effectively bind and guide AI agents – from the outside. In this episode, he shares with IT Ops Query his multi-faceted proposals for new industry standards in agentic IT governance and certification. But now that he's built it, will industry stakeholders come? 

Featuring: Jim Ford, founder and architect, Equilateral AI 

In today's episode, we'll cover: 

• The inherent conflict of interest in constitutional AI and guardrails 

• Authority drift and salience delusion 

• Proposed industry standard approaches to guiding and capturing agentic intent 

• How to avoid the pitfalls of SBOM 

And more! 

References:  

• The AI agent governance gap: How CIOs can gain control  

• AgenticOps for infrastructure faces a deterministic dilemma 

• Is AI our agent, or are our governments becoming agents for AI? | Computer Weekly 

• Authority Must Live Outside the Model – Equilateral 

To learn more about software development and platform engineering, check out Search IT Operations. 

To watch clips from our podcast, subscribe to our YouTube channel, @EyeOnTech.