@inproceedings{ziheng-etal-2026-simple,
    title = "Simple Role Assignment is Extraordinarily Effective for Safety Alignment",
    author = "Ziheng, Zhou  and
      Ding, Jiakun  and
      Zhang, Zhaowei  and
      Gao, Ruosen  and
      Wu, Ying Nian  and
      Terzopoulos, Demetri  and
      Kang, Yipeng  and
      Zhong, Fangwei  and
      Wang, Junqi",
    editor = "Liakata, Maria  and
      Moreira, Viviane P.  and
      Zhang, Jiajun  and
      Jurgens, David",
    booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {ACL} 2026",
    month = jul,
    year = "2026",
    address = "San Diego, California, United States",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-acl/2026.findings-acl.1164/",
    pages = "23249--23267",
    ISBN = "979-8-89176-395-1",
    abstract = "Principle-based alignment often lacks context sensitivity and completeness. Grounded in Theory of Mind, we propose role conditioning as a compact alternative: social roles (e.g., mother, judge) implicitly encode both values and the cognitive schemas required to apply them. We introduce a training-free pipeline featuring a role-conditioned generator and iterative role-based critics for refinement. Across five model families, our approach consistently outperforms principle-based, Chain-of-Thought (CoT) and other baselines across benchmarks. Notably, it reduces unsafe outputs on the WildJailbreak benchmark from 81.4{\%} to 3.6{\%} with DeepSeek-V3. Not only for common safety benchmarks, it consistently applies for agentic safety tasks. These results establish role assignment as a powerful, interpretable paradigm for AI alignment and LLM-as-a-Judge construction."
}