Artificial Intelligence: Hardwired to be a Colleague

Published: March 25, 2019

CybersecurityDEFENSEDARPAHHSDHSInformation TechnologyInformation TechnologyInnovationOMBProfessional ServicesResearch & DevelopmentResearch and DevelopmentVA

AI can transform the public sector by automating tasks and adopting innovative solutions by using advanced technologies to allow the government to shift from ‘low value’ to ‘high value’ tasks.

Artificial Intelligence is no longer relegated to technologists in R&D offices and quirky characters in sci-fi movies. Instead, AI is taking its place next to everyone in the public sector—from warfighters to auditors.

“The machines DARPA envisions will function more as colleagues than as tools,” Dr. Peter Highnam, deputy director of the Defense Advanced Research and Development Agency, told Senators recently.

Where AI began

Experts’ opinions vary on where AI began. It’s been suggested that the Tin Man in The Wizard of Oz helped to trigger the idea of robots thinking like humans. More specifically though, the field of AI can be traced back as far as a workshop held in 1956 at Dartmouth College. The workshop explored how machines can simulate human intelligence. Since then, there have been three waves of AI technology advancement.

First, experts took their knowledge of a particular area—the tax code, for instance—and characterized the data into computable rules. Tax-preparation software was a byproduct. The second wave is based on statistical learning. By inputting massive amounts of datasets, a system can perceive the natural world and predict situational outcomes. For instance, AI can assist a doctor make a diagnosis and choose the appropriate treatment for a patient.

Terms to Learn

Automation: Performing high-volume, repeatable tasks that humans normally perform
Machine Learning: Getting a computer to act without programming, such as the automation of predictive analytics
Supervised Learning: Labeling data sets so patterns can be detected and used to label new data sets
Machine Vision: Allowing computers to “see,” capturing and analyzing visual information via a camera, analog-to-digital conversion and signal processing.

The third wave combines the first two and adds contextual sophistication, abstraction and explanation. In this phase, a system would be built around a contextual model, be able to adapt to changes, and then explain its decisions. Think of the droid C3-PO, who is fluent in more than six million forms of communication and has a rationale for his decisions.

‘Innovates at the Speed of Relevance’

While the third wave of AI is in its infancy, the Department of Defense has put the first two waves to good use. In the FY2019 Budget Proposal, it intends to continue as one that “innovates at the speed of relevance.” The budget requests more than $84 billion in research, engineering and prototyping in areas that include AI and autonomous systems.

In the 2018 National Defense Strategy, defense officials plan to invest in the military application of autonomy, AI, and machine learning—including rapid application of commercial breakthroughs—to gain competitive military advantages. DoD aims to deliver AI-enabled capabilities that address key missions while using AI across the department through decentralized development and experimentation.

One example is the Sea Hunter. Beginning in 2014, DARPA and the Office of Naval Research collaborated to create the autonomous naval vessel. If AI prevails, the Sea Hunter may be the first of an entirely new class of vessel able to travel thousands of miles over the ocean for months at a time with no crew members aboard.

Indeed, “nations that are best able to adapt and integrate new technologies—in order to create speed and surprise across multiple domains in the fight—would prevail,” the FY2019 budget states.

Emerging from R&D Offices

AI isn’t limited to ships and satellites though. Civilian departments, including the departments of Health and Human Services, Veterans Affairs, and Transportation, are procuring AI solutions.

HHS’ Program Support Center (PSC) is looking for Intelligent Automation/Artificial Intelligence (IAAI) solutions to predict fraudulent transactions and identify critical suspects via facial recognition.

“PSC believes that IAAI solutions will be doing everything from reducing backlog and cutting costs to performing functions,” according to a Solicitation released on January 10, 2019. “IAAI will change the nature of many jobs and revolutionize facets of government operations.” 

Primary Requirements for AI (Tracked Opportunities, Civilian and Defense Agencies)
• R&D 43%
• IT 35%
• IT Services 24%
• Professional Services 18%
Source: Deltek

The Veterans Affairs Department wants to improve its interaction with the more than 9 million veterans it serves. It hopes to gain unbiased insights from analysis of the interactions between veterans, their families, and the VA’s 9,300 call agents. Through AI and machine learning, VA officials can discover customer patterns, predict or anticipate customer needs and augment agent tasks. AI-based robotic process automation (RPA) may eliminate manual tasks from the customer and agents.

Opportunities for AI in Civilian Agencies

Treasury Department, IRS: Advanced Analytics Artificial Intelligence and Machine Learning Capabilities for the Cybersecurity Cloud Solution Program
Transportation Department: Autonomous Unmanned Ariel Systems Research 
Veterans Affairs Department: Veteran Experience Data Curation Solution 
FBI: AI for Print Detection and Matching Next Generation Identification

AI can “transform the public sector by automating tasks and adopting innovative solutions utilizing advanced technologies enabling the government to shift from ‘low value’ to ‘high value’ tasks,” PSC predicts.

What is its future

The future of AI is somewhat foggy because it’s an evolving field. The Government Accountability Office hosted a forum in 2017 to hear about AI from industry, government, academia and nonprofit organizations. Participants agreed AI will bring benefits that cannot be predicted now and are hard to imagine. But, for sure, AI will improve economic outcomes and boost productivity. It will augment human decision-making and provide insights into complex and pressing problems. Areas that will certainly see the implications of AI developments include cybersecurity, automated vehicles, criminal justice, and financial services.

Despite the fog, the future is bright for AI and advances in technology, as federal officials recognize their benefits and have become more willing to invest today for benefits tomorrow.

However, AI will have its impacts. In its Solicitation, PSC adds that “IAAI technologies will fundamentally transform how the public sector gets work done—redesigning jobs and creating entirely new professions.”

How fundamental and transformative? The answer raises concerns among experts and officials.

The next blog post will dive into the impacts of AI and the need for policies to curb problems and address challenges.