The Regulatory Institute has launched a free reference corpus aimed at improving the substantive quality of AI-assisted legislative drafting.
The corpus contains more than 1,000 regulatory elements covering over 100 sectors and 80 jurisdictions. Designed for integration into retrieval-augmented generation (RAG) systems, fine-tuning datasets, and AI prompting frameworks, it helps identify missing regulatory elements and supports completeness checks in legislative drafting.
By addressing the substantive dimension of regulation, often overlooked by current legislative AI tools, the initiative seeks to enhance both AI-assisted drafting and regulatory capacity-building.
Further information and access to the corpus are available at www.howtoregulate.org/large-language-models-use/.