The Shape of the Cage: Deployment Geometry as an Under-Studied Variable in AI Safety

Zenodo (CERN European Organization for Nuclear Research) 2026 Preprint Open access

Read full text
Authors: Anthony W. Eckert
DOI: 10.5281/zenodo.18619444
Cited by: 0

Abstract

Application of the Void Framework to AI safety. Demonstrates that deployment geometry (how AI is presented to users) predicts harm better than model properties. Shows 9.4x vocabulary anomaly in AI researchers (p less than 0.001) and 0 percent drift in grounded agents versus 26 percent ungrounded default.

Fields of Study & Subjects

Ethics and Social Impacts of AI Safety Systems Engineering in Autonomy Explainable Artificial Intelligence (XAI)

Publication details

Journal / source
Zenodo (CERN European Organization for Nuclear Research)
Publication date
12 February 2026
Language
EN
Document type
Preprint