By Dennis D. McDonald

The following is how I introduced myself to the NIST Generative AI Public Working Group (NIST GAI-PWG) which I recently joined as a volunteer:

MY INTEREST

My background in consulting and project management related to IT and data related systems has influenced my interest in large langiage model based tools and their management.

A challenge to developing a risk management framework for generative AI is understanding how the use of these new tools doesn’t just allow fewer people to do more stuff better faster and cheaper, the tools can fundamentally alter how business processes are performed. As AI becomes embedded in how processes are managed they will fundamentally change how decisions are made and by whom.

This is fundamentally different from the changes we went through when automating low level and repetitive business processes. How those processes were managed didn’t fundamentally change how outcomes were managed. The same may not be true when automating white collar jobs involving judgement, prediction, and creativity.

While it’s tempting to control risk by making sure humans are inserted at key points throughout the system, that does not change that how we measure and respond to risk may also have to evolve just as business processes, altered by AI tool use, evolve.

SOME DOCUMENTS

Links to AI related documents I’ve published on my own web site that address some of these topics are here:

AI Governance