M.F.Borman
Making AI earn its place in regulated enterprises
Series #
Bounded Autonomy — what a boundary is actually made of, and why only part of it can be handed to a machine. Six posts, read in order.
Recent
Twin-datacentre continuity was built for accidents, not coordinated attacks
·4 mins
Twin-site designs assume one datacentre fails and the other keeps running. A coordinated attack does not play along.
An AI agent can write the change. It cannot allow it.
·9 mins
Contract theory tells you whether an AI agent’s change is correct. It never had to ask the second question: was the agent permitted to make it?
Rent the capability, own the control
·6 mins
Verifier-gated code generation looks cheaper. In a regulated enterprise it is a control architecture that produces the evidence you already owe.
Learned, then ratified
·7 mins
A learned system can tell you what your software did. It cannot tell you who is answerable for what that behaviour now obliges you to do.
AI moves what you audit, from the decision to the support behind it
·2 mins
AI cannot take the four judgements behind a boundary. It can assemble the menu of decisions you choose from. So the thing you audit moves up a level: from whether the decision is right to whether the support behind it was complete.