Compliance-by-Construction Argument Graphs: Using Generative AI to Produce Evidence-Linked Formal Arguments for Certification-Grade Accountability

Researchers propose a compliance-by-construction architecture that integrates Generative AI with structured formal argument representations to ensure accountability in high-stakes decision systems. The approach uses typed Argument Graphs, retrieval-augmented generation, validation constraints, and provenance ledgers to prevent AI hallucinations while maintaining traceability for regulatory compliance.

• →New architecture combines GenAI with formal argument structures to ensure every AI-assisted decision claim is supported by verifiable evidence.
• →System uses four key components: typed Argument Graphs, retrieval-augmented generation, validation kernel, and W3C PROV-aligned provenance ledger.
• →Approach addresses critical risks in current GenAI deployments including hallucinated reasoning, unsupported claims, and weak traceability.
• →Deterministic validation rules prevent unsupported claims from entering official decision records while allowing AI to accelerate argument construction.
• →Framework enables certification-grade accountability for AI systems used in safety-critical and regulatory environments.