Key Details
The brief argues that AI can make regulatory review faster and more systematic, but effective de-sludging depends on structured regulatory data, institutional ownership, and accountable human judgement.
Core Concept: The brief examines “policy sludge”—the accumulation of overlapping, outdated, inconsistent, and redundant regulatory requirements.
AI Application: Artificial intelligence can be used to identify duplicated reporting obligations, conflicting rules, and obsolete regulatory provisions across agencies.
Pilot Evidence: Municipal-level experiments found that AI-assisted reviews could identify opportunities to remove or consolidate more than 30% of reporting requirements.
Implementation Requirement: Effective AI review depends on machine-readable regulations, structured data repositories, and standardized metadata.
Governance Principle: AI should support regulatory review, but final decisions must remain with human regulators and policymakers.
Long-Term Approach: Regulatory review should become a continuous governance process, not a one-time reform exercise.
Summary
Regulatory Sludge as a Governance Problem
ADB Brief No. 394, “Cleaning Policy Sludge: Using Artificial Intelligence for Regulatory Review,” examines how regulations accumulate over time and create unnecessary friction for businesses, citizens, and public agencies. The brief defines policy sludge as the build-up of duplicated reporting requirements, outdated rules, inconsistent obligations, and overlapping compliance processes.
The report highlights that this accumulation happens in multiple ways. It can build vertically, from laws down to departmental circulars; horizontally, across agencies and jurisdictions; and temporally, as older provisions remain in force even after new rules are introduced.
How AI Can Support Regulatory Review
The brief argues that AI-assisted review can help governments examine large volumes of regulatory text more efficiently than manual review alone. Tools based on natural language processing, semantic similarity, and text-mapping can identify repeated obligations, similar reporting requirements, conflicting provisions, and areas where consolidation may be possible.
Evidence from pilot exercises suggests that AI-supported regulatory review can identify opportunities to remove or consolidate more than 30% of reporting requirements. The brief, however, treats AI as a decision-support tool, not a substitute for regulatory judgement.
Data Quality and Human Oversight Remain Essential
A central message of the brief is that AI-based de-sludging works only when regulations are available in machine-readable, structured, and well-maintained formats. Governments therefore need digital regulatory repositories, metadata standards, and clear institutional processes before automated review can be effective.
The brief also stresses the importance of human-in-the-loop governance. AI can flag possible duplication or redundancy, but final decisions on whether to retain, amend, consolidate, or remove a rule must remain with accountable public institutions.
From One-Time Clean-Up to Continuous Review
ADB argues that regulatory clean-up should not be treated as a one-off reform drive. Instead, de-sludging should be built into the regulatory lifecycle through pre-rulemaking checks, periodic reviews, and continuous monitoring of compliance obligations.
For India, this has direct relevance to ease-of-doing-business reforms, state-level single-window systems, regulatory guillotine exercises, and efforts to reduce compliance burdens across departments and jurisdictions.
What is Policy Sludge?
Policy sludge refers to unnecessary administrative friction created by excessive, outdated, duplicated, or poorly coordinated rules and procedures. In practice, it can appear as repeated form submissions, overlapping licences, unclear documentation requirements, or conflicting instructions from different agencies.
Reducing policy sludge can lower compliance costs, improve regulatory clarity, and allow public agencies to focus more on risk-based enforcement rather than routine paperwork.
Policy Relevance
Reduces Compliance Burdens for Firms and Citizens: AI-assisted review can help identify duplicated forms, repeated data requirements, and overlapping approvals that increase the cost of compliance, especially for smaller firms with limited administrative capacity.
Targets Cross-Agency and Cross-Jurisdictional Friction: The brief’s distinction between vertical, horizontal, and temporal sludge is especially relevant for India, where businesses often deal with overlapping requirements across central, state, municipal, and sectoral regulators.
Strengthens Ease-of-Doing-Business Reform: AI-based regulatory review can complement existing efforts by DPIIT, MCA, and state governments to simplify compliance, rationalise approvals, and reduce the number of low-value reporting obligations.
Builds the Case for Machine-Readable Regulatory Systems: The brief shows that meaningful AI-assisted review requires regulations to be digitised, searchable, structured, and supported by clear metadata. This makes regulatory data infrastructure a core part of administrative reform.
Preserves Accountability Through Human Oversight: The most important safeguard is that AI should identify possible sludge, not decide legal changes. Human regulators must determine whether an overlap is redundant, intentional, legally necessary, or politically sensitive.
Follow the Full Report Here: Asian Development Bank: Economics Publication Brief No. 394 — Strategy Guidelines on Capitalizing Artificial Intelligence to Erase Administrative Policy Sludge within Developing Economies