TY  - RPRT
TI  - Detecting and Preventing Harmful Behaviors in AI Companions: Development and Evaluation of the SHIELD Supervisory System
AU  - Ziv Ben-Zion
AU  - Paul Raffelhüschen
AU  - Max Zettl
AU  - Antonia Lüönd
AU  - Achim Burrer
AU  - Philipp Homan
AU  - Tobias R Spiller
PY  - 2025
UR  - https://arxiv.org/abs/2510.15891
ID  - 2510.15891
ER  -