More signs pointing to AI's growth in the federal market

Last week’s White House summit on artificial intelligence (AI) is an encouraging sign of American government and industry working collaboratively to advance this transformative technology. America’s defense leadership also grasps the critical importance of harnessing the power of AI for national security.

Defense Secretary Jim Mattis recently told a congressional committee that the Department of Defense (DoD) is “not going to have more papers, we’re going to move on [AI].” DoD is broadly pursuing AI, not just as another set of programs, but also as a powerful enabler for nearly every defense mission and function.

Strategic competitors are not standing idly by, either, as they reshape their economies to more service-based industries bolstered by technology. The U.S. commercial sector has a sense of urgency in adopting AI in the face of increasing international competition.

Federal civilian agencies need a similar sense of urgency when it comes to leveraging AI and machine learning to improve performance in their missions and functions. I am hopeful that the White House summit will spur efforts to bring more innovation like AI into agencies across the federal government. Fortunately, there is a recent model for this type of effort.

Last year’s “Report to the President on IT Modernization” and this spring’s President’s Management Agenda emphasize leveraging information technology (IT) and data for better service delivery.

One example is the Agriculture Department, an early test bed for new federal IT modernization Centers of Excellence. In addition, Congress appropriated money for the Technology Modernization Fund, a bold initiative to modernize legacy systems and reap savings as well as better performance.

Upgrading aging IT systems across the federal government is a crucial step, but it’s only a beginning.

As agencies continue to put in place more modern IT architectures and systems, such as cloud computing infrastructure, they can better leverage new innovations to improve performance and deliver better services to citizens. Perhaps most importantly, private sector commercial research and development (R&D) has led to amazing technology development and advances in AI. It is essential for government to harness the innovation that the private sector has to offer.

Today, U.S. and global commercial research spending is 25 times the government’s investment in R&D. Our military and civilian agencies need rapid access to innovation from government, contractor, and commercial sources.

For AI and other emerging technologies, government agencies can leverage and build on capabilities already developed in the private sector. No matter who develops AI, for what uses, what matters is improving government operations and staying ahead of our competitors. Contractors play essential roles in developing and delivering those capabilities to government agencies across the board.

How does AI fit into this need to access innovation? What type of AI should be used, and what are its potential applications?

Let’s start with a recent GAO report’s description of three types, or “waves,” of AI. First-wave AI comprises rules-based systems where a computer is programmed and produces outputs consistent with its programming, such as tax preparation or logistics scheduling software.

Second-wave AI systems, or machine learning, begin with data (including the outputs of first-wave AI) and use computing power to infer rules and then accurately predict outcomes based on data provided. When mature, autonomous vehicles and facial recognition programs will be good examples.

Third-wave AI systems are the future, and as the GAO report notes, they will be capable of contextual sophistication, abstraction, and explanation. Think of the computer HAL in the movie 2001: A Space Odyssey. The federal government has an opportunity to move quickly on incorporating first- and second-wave AI systems into improving government missions.

Applications of AI and machine learning have been underway across government for several years. The Department of Veterans Affairs uses AI to better predict medical complications and improve treatment for combat wounds at Walter Reed Medical Center. The Department of Homeland Security (DHS) has piloted AI tools for detecting cyber network intrusions and malicious activities. The Defense Advanced Research Projects Agency developed a digital tutor for computer skills learning that led to unprecedented performance by Navy students in training assessments.

The Bureau of Labor Statistics (BLS) is using AI to help analyze data for workplace injuries faster, doing in one day what would take a human alone one month to do. AI tools are becoming increasingly important for agencies to deliver citizen services and a high-performance government.

The President’s Management Agenda calls for using automation software to improve efficiency of government services. Many agencies have opportunities in areas where machine learning and AI tools hold great promise, from combatting improper payments to deploying agency assets effectively in response to natural disasters.

Yet, as the 2016 National Science and Technology Council report on AI noted, there is wide variance across agencies in the government’s capacity to foster and harness innovation. While some agencies have significant R&D budgets, a workforce with many scientists and engineers, and ongoing collaborations with private-sector innovators, many lack these attributes.

Bringing more of this type of innovation from America’s technology sector into our government will require strong leadership and innovative use of available contractor support. The White House should emphasize AI in government and make this an explicit focus area for the new Select Committee on Artificial Intelligence announced last week.

The White House could also accelerate AI in government by issuing a strategic vision and then providing implementation-level guidance for AI efforts across federal agencies. Incentivizing data sharing as part of that implementation will be critical for AI systems.

White House leadership and engagement will also be required for federal policy updates, including AI technical and governance standards that will need to evolve rapidly to ensure federal adoption of AI technologies. In addition, while AI technologies may augment or displace some types of work performed by federal employees, AI tools will enhance—rather than replace—the role of many workers.

Leveraging emerging technologies such as AI is the logical step following the White House’s strong emphasis on IT modernization. Ultimately, it is not replacing an old legacy computer with a newer one that delivers innovation that matters. It is the ability to leverage new tools—with the critical help of the private sector—that will help agencies redesign business processes that improve service delivery to the American people.