I have spent 10 years in AI and Compliance/Laws.. Here is what nobody warned me about !
I am here because i kept having this conversations again and again with engineer's, project managers, clients and so many more.
Engineers shipping ai into regulated workflows, founders building on top of LLMs and compliance teams absolutely doing the worse jobs governing something that moves faster then any framework they have dealt with..
The conversation, Always goes in the same direction.
Someone builds something smart, ground breaking, the next big thing and exactly six weeks later they are in a room with a auditor or a regulator, and thats the moment. they realize the gab between "A working model" and "A Complaint Model". The gap is wider then most teams realize.
Trust me, i know, i ve been on both sides of the table, I ve been a regulator and been the one who is being regulated.
Lets focus on the real stuff. ECOA adverse action trails. SR 11-7 model risk documentation.The EU AI Act and what high-risk actually means, California AI Act, Colorado AI Act, Shadow AI governance, HIPAA auditor requirements vs what engineers think the requirements are.
NO fluff, NO fear marketing. Connecting the people who are actually building something real.
So let me start by asking you guys.
What's the compliance or governance problem in your AI stack that you havent fully fixed yet ?
the one that may cause problems when a regulator finds you.
Drop it in the comments, Let's actually talk about it..