The Future Has Begun: Using Artificial Intelligence to Transform Government

Government agencies turn to artificial intelligence to buy equipment for warfighters, make American workplaces safer and much more.

By Peter Kamocsai, Associate Manager, Partnership for Public Service

Google held an event earlier this week to demonstrate an artificial intelligence-powered voice service by having it call a hair salon for an appointment. The Google Duplex feature used technology that sounded like human speech, including natural-sounding instances of “um” and “mm-hmm.” It’s possible the salon employee didn’t even realize she was conversing with software rather than a person.

AI is top of mind for many technology companies, as they introduce advances in the field. But government has a toe in the AI waters as well.

The White House is hosting an artificial intelligence event today with academia, government employees and 40 major U.S. companies. The administration plans to discuss how government can adapt government regulations for this game-changing technology, and advance AI research and development in the private sector.

Federal agencies were benefiting from AI long before the White House summit. The first AI contract was awarded in 1985 by the Social Security Administration, according to the federal government contracting database. The U.S. Postal Service was another early adopter, and uses the technology to read addresses on letters.

Earlier this year, the Partnership for Public Service and the IBM Center for The Business of Government, looked at four organizations using AI for the public good — two federal agencies, a local government and a research group supported in part by two federal R&D agencies. We explored the potential of AI to transform government and the way it does its work.

Many agencies started using AI years ago, and their work is now coming to fruition.

For example, the Department of Labor is using AI to improve efficiency in analyzing and understanding workplace injuries.

Since 2014, computers have been reading and analyzing a growing share of responses to Labor’s Survey of Occupational Injuries and Illnesses, an annual examination of workplace injuries. The computers not only can read responses faster than a human — a computer can do in one day what it would take a human one month to do — but also can analyze responses more accurately, on average.

The analysis is done through coding. Code words are assigned to pieces of information in a survey response by assigning, for example, the code “leg injury” to a narrative that describes that injury.

The potential benefits are substantial. Better coding leads to better analysis of workplace injuries, which means steps can be take to prevent injuries from happening again. This benefits workers, of course, but also the economy overall because fewer work hours are lost.

The Department of the Air Force, also featured in our report, plans to unveil an AI-driven website and phone app this year that helps acquisition professionals and contractors untangle the intricate web of acquisition regulations.

When the pilot program is finished, an Air Force acquisition professional or outside contractor will be able to ask the AI system about a specific purchase or contract, the way consumers might type a question into a web browser. The AI system will respond with relevant passages from acquisition regulations, rules, laws, past contracts, requests for proposals or training guides.

The potential benefits are huge in this case, too. The Air Force is one of the largest federal purchasers: In 2017 alone, it spent 11 cents of every dollar the federal government spent on acquisitions that year, totaling $53 billion. If the Air Force improves its acquisition activities, it could mean that life-saving new technologies get to warfighters faster or that startups can offer such technologies to the Air Force at all.

These examples are only the tip of the federal AI iceberg. Many other federal agencies use AI to do their jobs more efficiently, from the Central Intelligence Agency to the Environmental Protection Agency to the Internal Revenue Service. Others are looking at ways to incorporate the technology.

The Partnership’s report includes two other case studies. One is on a Kansas county using AI to help underserved populations get the services they need. The other is on how AI is helping wildlife rangers fight the menace of animal and plant poaching.

The report also presents issues and questions federal agencies should consider as they embark on the AI journey. Download the full report.