Category: A.I.
-
Preregistration: Temporal and stochastic instability of LLM-based risk ratings for IT general controls
Status: DRAFT — pre-data-collection protocol Version: 0.1 Authors: [Rozema, A.; co-authors TBD] Target venue: IEEE SaTML (backup: ACSAC, Computers & Security) 1. Background and motivation Commercial GRC products now sell LLM-generated risk rankings as a feature. Nobody has published whether those rankings are stable — across repeated runs, across model versions, across time — or…
-
A collection of Christmas cards that I am not allowed to send.
I’m sorry that you will not be receiving a Christmas card that looks like this. What follows is the progression of an AI conversation designed to develop holiday cards filled with, mirth, cheer, goodwill towards all, and that which is wonderful and beautiful in this the holiday season. What begins as joy and beauty, ends…