Page 10 - Federal Computer Week, March/April 2019

Artificial Intelligence
By automating rote tasks and facilitating better decision-making, artificial intelligence has implications for every aspect of government
LONG CONSIDERED THE stuff of science fiction, artificial intelligence is now very much a part of daily
life. Whether people are interacting with
a smart refrigerator at home or a chatbot on a government website, they are rarely surprised by AI capabilities these days and, in fact, have come to expect them.
When applied properly, the technology can boost the public sector’s efficiency and effectiveness. But AI is also a
key component of IT modernization
and innovation, and it promises to revolutionize citizen engagement — a top priority for many government leaders.
The use of AI isn’t new in the public sector. The government’s interest and research into the possibility of creating
a machine that thinks began in the 1950s. By the 1990s, AI technology was being used for activities as prosaic as recognizing handwritten addresses on letters so they could be automatically routed through the U.S. Postal Service
— saving hundreds of millions of dollars, according to a report by the Harvard Ash Center for Democratic Governance and Innovation.
Several technological innovations have hastened AI adoption in recent years.
Those innovations include the internet of things, data analytics, cloud computing and graphics processing units. In fact, GPUs can power many of the elements essential to AI, including robotic process automation and facial recognition. For instance, the supercomputer Summit
at the Energy Department’s Oak Ridge National Laboratory became the fastest in the world when it debuted last year with the ability to perform more than 3 exaops — or 3 billion billion — calculations per second, with 95 percent of that computing power coming from GPUs.
Successful AI implementations are having an impact at other civilian and Defense Department agencies. The Army’s Sgt. Star is a virtual guide to GoArmy. com that handles the work of 55 human recruiters; it uses machine learning to improve its ability to answer questions from potential recruits. The Agriculture Department’s National Institute of Food and Agriculture applies AI to help farm, forest and ranch managers make decisions that support sustainability.
The Department of Homeland
Security and NASA’s Jet Propulsion Laboratory have partnered to develop AUDREY — which stands for Assistant for Understanding Data through Reasoning,
Extraction and Synthesis. The system synthesizes high-level data at the scene of an emergency to glean insights that help first responders make better decisions. AUDREY will be put to the test in Canada this year.
State and local governments have
also set their sights on AI. Vermont,
for instance, is using the technology to monitor bridge deterioration and predict with 85 percent accuracy how long road treatments will last.
In addition, when the National Association of State CIOs conducted its 2018 survey, 57 percent of its members named AI as the emerging IT that will have the biggest impact in the next three to five years.
Expanding investments in AI
Two components of AI are driving innovation today. Machine learning is based on algorithms that learn from data to make decisions or predictions. Deep learning mimics humans’ neural networks and is
an approach to implementing machine learning.
However, no single definition exists
for AI as a whole. In fact, the fiscal 2019 National Defense Authorization Act directs the Defense secretary to produce a
S-10 SPONSORED CONTENT
Liu zishan/Shutterstock/FCW Staff