Personalized Constitutionally-Aligned Agentic Superego: Secure AI Behavior Aligned to Diverse Human Values
Information · 30.07.2025 · 16 (8) · 651
DOI: 10.3390/info16080651 ↗Kurz gefasst
Fachartikel in der Zeitschrift Information, online erschienen am 30.07.2025. Verfasst von Nell Watson und 4 weiteren Autor:innen. Die Arbeit wurde bisher 2-mal zitiert (Crossref) und nennt 49 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: The Future of Life Institute, AI Safety Camp, The Survival & Flourishing Fund.
Abstract (Originalsprache)
Agentic AI systems, possessing capabilities for autonomous planning and action, show great potential across diverse domains. However, their practical deployment is hindered by challenges in aligning their behavior with varied human values, complex safety requirements, and specific compliance needs. Existing alignment methodologies often falter when faced with the complex task of providing personalized context without inducing confabulation or operational inefficiencies. This paper introduces a novel solution: a ‘superego’ agent, designed as a personalized oversight mechanism for agentic AI. This system dynamically steers AI planning by referencing user-selected ‘Creed Constitutions’—encapsulating diverse rule sets—with adjustable adherence levels to fit non-negotiable values. A real-time compliance enforcer validates plans against these constitutions and a universal ethical floor before execution. We present a functional system, including a demonstration interface with a prototypical constitution-sharing portal, and successful integration with third-party models via the Model Context Protocol (MCP). Comprehensive benchmark evaluations (HarmBench, AgentHarm) demonstrate that our Superego agent dramatically reduces harmful outputs—achieving up to a 98.3% harm score reduction and near-perfect refusal rates (e.g., 100% with Claude Sonnet 4 on AgentHarm’s harmful set) for leading LLMs like Gemini 2.5 Flash and GPT-4o. This approach substantially simplifies personalized AI alignment, rendering agentic systems more reliably attuned to individual and cultural contexts, while also enabling substantial safety improvements.
Förderung
- The Future of Life Institute
- AI Safety Camp
- The Survival & Flourishing Fund
Zitieren
APA
Watson, N., Amer, A., Harris, E., Ravindra, P., & Zhang, S. (2025). Personalized Constitutionally-Aligned Agentic Superego: Secure AI Behavior Aligned to Diverse Human Values. Information, 16(8), 651. https://doi.org/10.3390/info16080651
Vancouver
Watson N, Amer A, Harris E, Ravindra P, Zhang S. Personalized Constitutionally-Aligned Agentic Superego: Secure AI Behavior Aligned to Diverse Human Values. Information. 2025;16(8):651. doi:10.3390/info16080651
BibTeX
@article{Watson2025,
author = {Watson, Nell and Amer, Ahmed and Harris, Evan and Ravindra, Preeti and Zhang, Shujun},
title = {Personalized Constitutionally-Aligned Agentic Superego: Secure AI Behavior Aligned to Diverse Human Values},
journal = {Information},
year = {2025},
volume = {16},
number = {8},
pages = {651},
doi = {10.3390/info16080651}
}
EUVIMED