How Artificial Intelligence Can Transform Government

At the risk of dating myself, one of my favorite movies growing up as a kid was "WarGames" starring Matthew Broderick. I didn’t realize it at the time, but in the climactic scene, the large supercomputer ‘WOPR’ operated by the Defense Department, showed artificial intelligence capabilities. By playing tic-tac-toe against itself, it learned a lesson that prevented global thermonuclear war.

In many ways, Hollywood has warped what many think of when they first hear the term artificial intelligence, or AI. My thoughts used to go to movies like "The Terminator" or "The Matrix" where sentient machines develop the ability to think for themselves and try to overthrow humankind. While this makes for an exciting movie plot, AI has much more tangible—and less threatening—benefits, particularly for government.

In 2018, U.S. Chief Information Officer Suzette Kent announced the creation of the first Federal Data Strategy that will serve as a foundation for how agencies use AI.

Her analogy in describing the need for the strategy was compelling.

“Technology modernization allows us the opportunity to rethink our foundation,” Kent said at an event announcing the strategy. “We have to move aggressively. We don’t want to build the high-speed train without the track.”

AI can serve as part of that track. As the government collects more and more data, the need for solutions to drive true value from that data grows in importance. AI, in conjunction with big data and analytics, can deliver that baseline value and go beyond traditional solutions to find deeper insights.

Other governments have recognized this as well. For example, the United Arab Emirates was the first nation to appoint a senior cabinet official solely focused on AI empowerment and oversight within the government, appointing a Minister of State for Artificial Intelligence in October 2017. Canada was the first nation to release a national AI strategy. And China has released a 3-year plan to be a leader … if not the leader … in AI.

Understanding AI

So, for those of us whose understanding of AI has heretofore been solely that of the Hollywood blockbuster, AI is the science of training systems to emulate specific human tasks through learning and automation. In short, it’s a technology that makes it possible for machines to learn from experience, adjust to new inputs and perform specific human tasks, such as pattern recognition, finding anomalies in data, image and video analytics, and more. Specific to analytics, AI can help analytics programs in government find connections and trends in the data that human analysts might miss due to scale, complexity, or other factors … and it can do it at a much faster speed. AI can find context in data, gaining insight from previous discoveries to create better outcomes in the future. From an analytics perspective, AI tends to focus in these areas:

• Machine learning: Machine learning and deep learning find insights hidden in data without explicitly being told where to look or what to conclude. This results in better, faster and more accurate decision-making capabilities.
• Natural Language Processing: NLP enables understanding, interaction and communication between humans and machines, automatically extracting insights and emerging trends from large amounts of structured and unstructured content.
• Computer vision: Computer vision analyzes and interprets what’s in an image or video through image processing, image recognition and object detection.
• Forecasting and optimization: Forecasting helps predict future outcomes, while optimization delivers the best results given resource constraints. This includes enabling large-scale automation for predicting outcomes and optimizing decisions.

Ultimately, this is all based on the idea that systems can learn from data and identify patterns to make decisions with minimal human intervention.

The Value Proposition for Government

What these capabilities do is make sense of large amounts of data quickly and efficiently. These insights can be used in almost every aspect of government from improving operations and management to determining the value of programs and finding underserved segments of the population that need additional assistance.

“We have a mandate to automate many processes now through law. We have a lot of automation tools that provide easy steps,” Kent said. “That’s what our citizens expect—instantaneous interaction and service. That’s not the future. Those are things we can do today.”

AI is already widely used in every sector of the economy. Private businesses use these solutions to find efficiencies in their own business and improve the return-on-investment of products and projects. Government agencies on all levels have begun to utilize it as well, using AI in a vast array of applications, including:

• Automatically detecting fraud and improper payments before money goes out the door, and ensuring that benefit agency funds and services go to those who really need them.
• Enabling public safety agencies to make sure the things citizens consume—like food and medicine—are safe.
• Empowering law enforcement and criminal justice agencies to improve efficiency and consistency across cases.

The Federal Data Strategy is incredibly important. Its goal is to build a strategy that will transform how the government operates and delivers services.

This will prove to be an important document and shows that the federal government understands the value that AI can bring. Investing in AI can provide immediate dividends and quickly drive improvements inside government agencies. Kent obviously understands the importance of these technologies, and with the Federal Data Strategy, agencies will find themselves on the right track.

Steve Bennett, Ph.D.,  is director of SAS' Global Government Practice. He is the former director of the National Biosurveillance Integration Center within the Department of Homeland Security.