What question did this study set out to answer?

This work proposes a new framework for the alignment of large language models (LLMs) with human-centered design principles.

May 6, 2026Open Access

View Full Paper

Rule-Governed Systems in Adaptive Environments: An HF-Grounded Framework for LLM Alignment

Key Points

This work proposes a new framework for the alignment of large language models (LLMs) with human-centered design principles.
Draft proposal for a new alignment framework focused on rule-governed systems.
Call for collaboration on integrating human factors and resilience engineering into LLM designs.
Highlights the architectural insufficiency of current LLM alignment frameworks for high-stakes interactions.
Suggests incorporating adaptive, context-aware, and error-tolerant principles into future LLM designs.

Abstract

A working draft, problem-space proposal, and call for collaboration:Contemporary large language model (LLM) alignment frameworks are predominantly structured around machine learning optimization, policy-layer rule sets, and compliance-oriented legal risk management. Such approaches are architecturally insufficient for systems designed for sustained, high-stakes interaction with human beings, because they do not incorporate the adaptive, context- aware, and error-tolerant design principles established by decades of research in human factors engineering, resilience engineering, and ecological psychology.

Perguntar à IA

Bookmark

View Full Paper

Perguntar à IA

Bookmark

View Full Paper

Rule-Governed Systems in Adaptive Environments: An HF-Grounded Framework for LLM Alignment

Key Points

Abstract

Cite This Study