AI to Scan Federal Rules Annually: Agencies Get 30 Days to Scrub 'Outdated' Regulations

The Leveraging Artificial Intelligence to Streamline the Code of Federal Regulations Act of 2025 is setting up the biggest digital spring cleaning the federal government has ever seen. Starting within 90 days, the Office of Management and Budget (OMB) must deploy an Artificial Intelligence (AI) system every year to scan the entire Code of Federal Regulations (CFR) and flag rules that are either “redundant” (doing the same job as another rule) or “outdated” (irrelevant due to new laws or tech).

This isn’t just about making the rulebook shorter; it’s about speeding up how rules are killed off. If the AI flags a regulation, the agency that created it gets just 30 days to decide its fate. If they agree it’s redundant or outdated, they have another 30 days to remove or update it. Crucially, the bill lets them skip the usual, lengthy administrative steps—like public notice and comment periods—required for changing or ditching federal rules. This gives agencies a massive shortcut to regulatory reduction, but it also means the public loses their voice in the process.

The Algorithm Is In Charge

For the AI system to be used, it has to meet strict standards set by the National Institute of Standards and Technology (NIST), covering everything from accuracy and transparency to national security risks (SEC. 3). This is a good thing, ensuring the tool isn't just some buggy software. However, the entire process hinges on how the AI defines “redundant” or “outdated.” Will it flag a rule protecting a specific wetland because a different, broader environmental rule exists? Will it see an old safety standard as irrelevant because newer technology could make it obsolete, even if that tech isn't universally adopted yet?

For the average person, this introduces a new layer of uncertainty. If you’re a small business owner relying on a clear, specific regulation about, say, material handling, that rule could disappear in 60 days if the AI flags it and the agency agrees. Since the agency's final determination is made quickly and is explicitly stated as “final” (SEC. 3), there’s very little time for outside groups to object or even notice before the rule is gone.

The Price of Speed: Bypassing Public Input

To achieve this fast-track removal, the bill amends existing law (5 U.S.C. 553(b)) to allow agencies to use an expedited process to remove flagged rules (SEC. 4). This means they bypass the standard administrative procedure, which is the mechanism that ensures accountability. Normally, when an agency wants to change a rule, they have to publish the proposed change and allow the public—you, your local community group, or a trade association—to weigh in. This bill removes that requirement entirely for AI-flagged rules.

While the benefit is clear—less regulatory bloat means less time wasted navigating unnecessary complexity—the risk is that rules designed to protect consumers, the environment, or worker safety could be swept away without public scrutiny. Imagine a specific food safety rule is flagged as redundant. If the agency decides to scrap it quickly, the public won't have the chance to argue why that specific rule is still necessary, perhaps because it covers a niche area the broader rules miss.

Transparency with a Catch

On the plus side, the bill requires agencies to immediately post their decisions—whether to update or remove a rule—on their website, along with a short written explanation (SEC. 3). This is a win for transparency. However, there’s a curious provision that allows the head of the agency to send a separate, classified document to Congress to add more detail to that public explanation. If the full rationale for removing a regulation is sent to Congress under wraps, it raises questions about whether the public is truly getting the whole story behind why a long-standing rule was suddenly deemed unnecessary.