The National Institute for Standards and Technology just delivered a final plan to the White House for developing artificial intelligence standards for the private sector, in order to “maintain and strengthen America’s leadership in AI.”

But the private sector seems to be developing standards just fine on its own. Where we really need to promote the use of AI (and establish AI standards) is actually within government, because in the few governmental bodies where AI has been adopted, the results have been remarkable.

AI is the use of computers to perform cognitive functions that we usually associate with humans, such as learning or problem-solving. AI can pick up on patterns, trends, similarities and differences in data; interpret images, videos and spoken languages; autonomously operate machines; and complete numerous other tasks that expand human potential.

Although the majority of AI applications occur in the private sector, the potential for revolutionary innovations exists within the public sector as well.

AI can be used in countless ways to improve processes within government or to provide better services for the public. Unfortunately, the adoption of AI within government has, at least so far, been somewhat limited.

Nonetheless, where government organizations have adopted AI, the results speak for themselves.

For example, the Cincinnati Fire Department used AI tools to predict whether an ambulance was needed in response to a 911 call.

This reduced delays in getting patients to the hospital by 22 percent. That means roughly 2,500 more people will receive the immediate medical care they need annually, due to just one city department taking one initial step with AI.

Yet AI has been adopted by federal government agencies on a larger scale, too — especially in applications related to law enforcement, national security and the military.

By 2016, for example, the Department of Veterans Affairs was using AI to help predict medical complications and better treat combat wounds, which improved patient outcomes and recovery times while reducing costs.

Better services aren’t the only benefit of adopting AI in the public sector. Government agencies could use AI to improve the quality of regulations as well.

AI could help agencies identify and improve regulations that may hinder innovation or productivity growth due to regulations’ complexity or due to uncertainty about how regulations will be enforced.

A 2016 Mercatus Center study found that regulatory accumulation since 1980 has reduced the size of the U.S. economy by $4 trillion, and this regulatory buildup disproportionately impacts groups like low-income households.

A good first step would be for our government to follow its own lead: Create a plan to establish standards for the adoption and development of AI — and apply those standards only to the public sector.

Simply establishing a government-wide standard for public documents would generate enormous benefits. Such a standard could require that all documents be machine-readable and include “tags” (similar to hashtags on social media) for particular elements of the documents that would allow users to sort by topic or specific content, or pull together related data from multiple documents.

This would allow both government and the private sector to aggregate, organize, filter and analyze latent data that often exist only in PDF format across hundreds of websites — assuming these documents are even available online.

Creating such a standard for regulations would allow anyone to easily access various information about all, some or even individual regulations.

Among other things, this would allow businesses to determine which regulations apply to them without scouring the nearly 200,000 pages of regulations currently on the books.

This would also help achieve another goal of the current administration: opening up government data for public use. And that includes government use: A modernized standard for regulation would let federal, state and local regulators learn from each other by providing insights into the effectiveness of different regulations, and even enable them to identify overlapping or contradictory regulations.

The government’s growing interest in AI is promising. And working with the private sector is important to ensure that agencies understand the complexities and potential safety concerns of this new technology. But there is also an opportunity to improve the lives of all Americans by leveraging AI within the government. If we’re devising plans to create a better future through AI, that’s where we should start.

Patrick McLaughlin is the director of policy analytics at the Mercatus Center at George Mason University. Tyler Richards is the research coordinator of policy analytics at the Mercatus Center.