Hitchhikers Guide to AI Governance Blog

It's not just having an AI system up and running, but what happens if the AI outputs cannot be trusted or are unsafe?, writes Bruce Mullan

The hardest part of automating invoice processing isn't teaching the AI to read them properly, it's fixing everything around them, writes Bruce Mullan

A botched Copilot AI deployment by an ASX listed company allowed unrestricted access to payroll data. Someone forgot to define the scope, purpose and boundaries of the AI system which is part and parcel of AI Governance.

We need to think more carefully about our back office business systems. Does your ERP system need a kill switch? More than likely it does, writes Bruce Mullan

How to build an AI system and use case inventory for effective AI risk management, writes Bruce Mullan

If you don't know where you're AI is landing, you'll probably end up somewhere else. Defining multiple success criteria is critical, writes Bruce Mullan

One employee ... one bad prompt ... one company-wide incident ... in one second. You can't govern what you can't see, writes Bruce Mullan

The trouble with "safe" computer systems is they exist in a virtual world. Safety-by-design helps reduce AI's non-physical harm capability writes Bruce Mullan

AI offers extraordinary Speed. It enables unprecedented Scale. With increasing organisational Sophistication it can operate Safely, writes Bruce Mullan