The following elements may be used to complement regulations with regard to sanctions and accompanying measures. The elements can be used both to establish sanctions that apply automatically by virtue of the law and to give authorities or courts/tribunals discretionary powers.
Click here to download the list of sanctions.
The list was in the first round drawn up on the basis of model laws developed by the Regulatory Institute (www.howtoregulate.org) because these model laws integrate regulatory knowledge from many jurisdictions and sectors alike. They therefore provide an excellent international and cross-sectoral basis. It was entirely drawn up by various artificial intelligence programmes, known as Large Language Models (LLMs)1. It is the result of a long series of instructions and tasks that took the human in charge several days to complete, whilst the document was only marginally revised by humans.
Namely the LLM interface Perplexity2 has in many cases suggested additional very sensible rules based on regulations that have been sifted through around the world. In doing so, Perplexity continued the Regulatory Institute’s approach of learning from regulators in all sectors around the world and making the knowledge gathered available globally and across sectors. It is this approach that has been continued by regulatory practitioners supporting the Regulatory Institute who manually redacted the text to a modest extent. Finally, it has been presented to members of the artificial intelligence working group of the Commonwealth Association of Legislative Counsel. The document is therefore the result of four rounds of international gathering of regulatory knowledge. Feel free to suggest further improvements.
We asked the LLMs to provide their references. The elements, and therefore the references, come from legislation all over the world. The letters “ML” stand for “Model Law of the Regulatory Institute”. You can therefore find the relevant model law provisions via this website: https://www.howtoregulate.org/category/ri-model-laws/.
The list overlaps slightly with the parallel list of powers and obligations which is recommended as an extension for cases where this list of sanctions and accompanying measures is not considered sufficient3. The difference between the two lists is that this list pursues the goal of compliance only indirectly via sanctions, whereas the list of powers and obligations is aiming at directly ensuring compliance. Both lists complement the document “Cross-sectoral standard provisions for regulation“, which already contains some basic sanctions and accompanying measures. The document “Cross-sectoral standard provisions for regulation” also contains an Annex / Chapter dealing with procedural and side-aspects of civil / administrative sanctions. Hence the two documents should be read together, they go hand in hand.
Despite many deletions, the list might still contain some redundancies. We also feel that the order of the elements could still be improved. Finally, we are concerned that some of the references might not be correct or do not contain a sanction or accompanying measure, but only an obligation that could also be imposed as a sanction. Despite these flaws, the list can already be used to make draft legislation more complete in terms of sanctions. This is the list’s sole purpose.
1 By far the best performing LLM for extracting elements from the Regulatory Institute’s model laws was Claude Sonnet 3.5 (Anthropic). By far the best performing LLM for completing the list with examples from the internet was Perplexity. However, some contributions were made by Chat GPT 4o, Deepai and Mistral.
2 Strictly speaking, Perplexity is not an independent LLM, but an interface that builds on various LLMs.
3 In theory, any kind of obligation can be imposed to sanction a person