Argumentation for Explainable and Globally Contestable Decision Support with LLMs

Researchers introduce ArgEval, a new framework that enhances Large Language Model decision-making through structured argumentation and global contestability. Unlike previous approaches limited to binary choices and local corrections, ArgEval maps entire decision spaces and builds reusable argumentation frameworks that can be globally modified to prevent repeated mistakes.

• →ArgEval addresses LLM opacity issues by providing structured evaluation of general decision options rather than instance-specific reasoning.
• →The framework builds task-specific decision space maps and corresponding option ontologies for more comprehensive analysis.
• →Unlike existing methods, ArgEval supports global contestability that modifies underlying decision logic to prevent repeated errors.
• →Initial testing on glioblastoma treatment recommendations shows the framework can produce clinically aligned explainable guidance.
• →This approach shifts from post-hoc reasoning to proactive structured evaluation of decision frameworks.