Government agencies play a pivotal role in crafting the rules and regulations that bring primary law – emanating from legislatures and parliaments – to life.
The volume of rules and regulations has grown exponentially in many jurisdictions over the last few decades, with the total number of pages published in, for example, the US Code of Federal Regulations ballooning from 50,000 to more than 180,000 pages. This expansion translates to increased regulatory burdens for businesses and heavier administrative burdens for agencies.
Any content library of that size is going to contain redundancies and overlaps. A natural question to ask is: “How much of any library of regulations contains overlapping material?” A natural follow-on question is: “What factors contribute to the creation of the overlap?”
Getting to the root causes of regulatory overlaps
The rapid increase in regulatory footprint since the 1970s has likely been exacerbated by the use of computers in the drafting process. The ubiquitous ability to copy and paste has enabled effortless duplication of existing regulatory text as the first part of a copy/paste/tweak pattern i.e. take regulatory language that is close to what you need. Copy/paste it. Then make some edits.
A second contributory factor is the distributed nature of the teams working on regulations in agencies. For example, a lawyer working on a road-related piece of regulation for a Department of Health might discover a similar regulation already issued by the Department of Transportation. It would be possible, in theory, to collaborate on a re-factoring of the existing regulation to cover the needs of both agencies but, unfortunately, siloed organizational structures can make this difficult in practice.
The volume of rules and regulations has grown exponentially in many jurisdictions over the last few decades.
Addressing regulatory overlaps
In 1980, the Paperwork Reduction Act was passed by the US Congress as a legislative tool to address regulatory burden. Similar laws exist in other jurisdictions. Some have an automatic sunset provision in each regulation, designed to reduce the likelihood of unnecessary regulations staying in place over the long term. Some have a mandatory review process whereby the legislature/parliament retains oversight on regulations throughout their life cycle. Some operate a ‘one-for-one’ rule (or similar) whereby an agency wishing to create a new regulation, must find an existing one that can be repealed.
The consequences of regulatory footprint expansion
1. Administrative burden
2. Impact on businesses
3. Economic impact
4. Government efficiency
Clearing the path: reducing overlap in the regulatory footprint with technology
Modern software tools, especially artificial intelligence (AI), offer opportunities to identify and remove duplications in even the largest libraries of rules and regulations.
AI can analyze vast regulatory databases, flagging duplicates or near-duplicates at a speed and scale that would have been impossible just a few years ago. This analysis can be used to de-duplicate the existing regulatory stock. It can also be built into drafting processes allowing drafters and administrative oversight bodies alike, to be notified of the presence of emerging duplication while the new regulatory language is still in draft form.
A data-centric approach to managing rules, regulations, and policies
To fully harness AI’s potential, agencies must shift to data-centric approaches to regulatory management. Indeed, by utilizing knowledge graphs in regulatory libraries, rule-makers can tap into the opportunities of AI but staying grounded in the authoritative master texts at all times.
This knowledge graph approach also increases efficiencies in managing the domino effect of change that amendments to laws and regulations can trigger. For example, a new or changed primary law may have knock-on impact to a regulation. A change to one regulation, may have a knock-on-effect to another regulation that cites it etc.
Knowledge graphs enable Large Language Models to operate with greater structure and context, reducing risks such as hallucinations while also provide rule-makers with the ability to ‘converse’ with their own content and deliver AI-rich insights downstream. Not only does this reduce the likelihood of redundancies, it improves the accessibility of rules for all stakeholders.
