Design Behaviour Codes (DBCs): A Taxonomy-Driven Layered Governance Benchmark for Large Language Models

Researchers introduce the Dynamic Behavioral Constraint (DBC) benchmark, a new governance framework for large language models that reduces AI risk exposure by 36.8% through structured behavioral controls applied at inference time. The system achieves high EU AI Act compliance scores and represents a model-agnostic approach to AI safety that can be audited and mapped to different jurisdictions.

• →DBC framework reduces AI model risk exposure rate from 7.19% to 4.55%, representing a 36.8% relative risk reduction.
• →The system is model-agnostic and can be applied at inference time without requiring model retraining.
• →EU AI Act compliance scoring reaches 8.5 out of 10 under the DBC governance layer.
• →The benchmark evaluates 30 risk domains across six clusters including bias, privacy, robustness, and malicious use.
• →Researchers released benchmark code and evaluation tools to enable reproducible AI safety research.