Page 11 - FCW, March/April 2020

Strategic Plan that identifies critical areas
in need of federal funding, the National Institute of Standards and Technology’s (NIST) release of the U.S. Leadership in
AI plan for developing technical standards, and the establishment of the National AI Institute at the Department of Veterans Affairs with the stated purpose of advancing the health and well-being of veterans.
In November, the National Security Commission on AI released its interim report to Congress, with its initial assessment of AI’s connection to national security, areas
for improvement and preliminary steps the government could take. The report also notes the broad, bipartisan support the commission has received across the public and private sectors.
As a result of all those efforts, the government is starting to allocate more money to AI initiatives. The fiscal 2020 budget provides about $850 million to support the White House’s American AI Initiative at four agencies — Energy Department, National Institutes of Health, NIST and National Science Foundation. The Networking and IT Research and Development Program issued a supplement to the budget that added AI as a program component area for the first time and made a non-defense budget request of $973.5 million for AI.
Beyond automating rote tasks
Many early AI efforts have focused on implementing chatbots to save employees’ time and increase citizen engagement. The Agriculture Department is developing a chatbot to help USDA customers more
easily find the information they need while freeing employees to concentrate on higher- level activities. State and local agencies are using chatbots to answer questions about government services such as license plate renewals and to report problems such as potholes.
AI’s capabilities extend far beyond answering users’ routine questions, of course. For instance, cities are adding sensors to traffic lights to better manage vehicles’ flow through congested areas — and thereby reduce traffic accidents, dangers to pedestrians and pollution. In addition, NASA’s Jet Propulsion Laboratory and the Department of Homeland Security collaborated on the Assistant for Understanding Data through Reasoning, Extraction and Synthesis (AUDREY)
to provide situational awareness for first responders at the scene of an emergency.
In a press release, DHS officials said AUDREY is personalized to the individual responder and “leverages human intelligence and collects data to achieve better machine intelligence and provides insight that first responders may not have in the crucial moments of an emergency.”
At the Defense Department, officials have also recognized the power of AI. They created the Joint AI Center in 2018 to accelerate the technology’s adoption across DOD. Although AI is not yet ready to tackle complex activities such as nuclear command and control or missile defense, it eventually will be, said Lt. Gen. Jack Shanahan, the center’s director, in an article published on Defense.gov last October. “There is no part of the Department of Defense that cannot
benefit from AI,” he added.
The Army is already using AI for
predictive maintenance on its aging Bradley Fighting Vehicle fleet and successfully predicted failure on a major subsystem two weeks after the capability launched, which avoided downtime and improved soldier safety. Last September, the Air Force released an AI strategy to “harness and wield the most representative forms of AI across all mission- sets, to better enable outcomes with greater speed and accuracy, while optimizing the abilities of each and every airman,” officials wrote in a press release.
What’s more, AI is finding an expanding role in cybersecurity, with many agencies looking for commercial tools that can detect and respond to incidents on its networks without human intervention. Along the
way, agencies also recognize that their employees need better training to deploy and manage AI tools. The Defense Acquisition University is one of them. Officials issued a request for information on solutions for an adaptive learning environment that responds to students in real time. In other words,
it wants to use AI to teach 174,000 DOD acquisition professionals about AI.
As AI adoption grows, challenges related to transparency, explainability and privacy are emerging. Many experts are urging the U.S. government to take a leadership role in addressing those challenges and ensuring the ethical use of AI worldwide. The technology has come a long way in recent years, and the stage is now set for AI to improve government — and people’s lives — in previously unimaginable ways.
28%
FCW survey respondents who said their agencies have begun deploying AI tools
30%
FCW survey respondents who said their teams have received some training in data science and AI
55%
States that are actively pursuing AI
84%
U.S. public-sector leaders who cited data privacy and quality as the biggest AI adoption challenges
44%
Estimated growth rate in AI spending by federal governments worldwide through 2022
Sources: Deloitte, FCW, IDC, National Association of State CIOs
SPONSORED CONTENT S-11