TY  - RPRT
TI  - Position: Safety and Fairness in Agentic AI Depend on Interaction Topology, Not on Model Scale or Alignment
AU  - Tanav Singh Bajaj
AU  - Nikhil Singh
AU  - Karan Anand
AU  - Eishkaran Singh
PY  - 2026
UR  - https://arxiv.org/abs/2605.01147
ID  - 2605.01147
ER  -