AI makes it possible to take a people-first approach to technology that adapts to the needs and characteristics of individuals and how we interact now in our digital world. In the past, people used to have to learn to think like computers and how they were programmed. Now, computers have learned to alter their processes in order to think like people. This is a very important strategic shift in federal government, as it presents an opportunity to rethink how agencies can use technology to better serve citizens—achieving more by allowing them to realize real outcomes, not simply transact with government in predefined ways. Up to 94 percent of business and IT executives in the public-service industry anticipate making extensive investments in one or more AI-related technologies over the next three years to accommodate this shift.1
Basic decisions that a human can make in under one second can now be automated using intelligent technology. Automation enables employees to focus on more complicated tasks, and enhances more complex cognitive processes by providing insights that inform better actions. Eighty percent of public service executives reported that the automation of certain simple tasks through machine learning will “free up employees to focus on more critical—and rewarding—activities that are more closely aligned with citizens’ needs,” and that this “will improve the job satisfaction of current employees,” thereby improving retention.2
With profound benefits to both the citizens they serve and the people they employ, federal leaders are embracing the advances of artificial intelligence. AI solutions are getting easier to implement and automate, but agencies need to think through all of the implications before deployment. This is truly a transformation of the relationship between humans and machines—it requires a thoughtful, and thorough, analysis.
For example, AI interfaces that interact directly with citizens must be held accountable in ways similar to person-to-person interactions. To avoid challenges in interactive modes, user-centered design activities aligned with AI interfaces should be included to ensure seamless, clear and secure communication dialogues between AI and people.
In cases where AIs are augmenting employees through the proactive presentation of relevant information, it is important to place that information in context, with links to backup materials and evidence that supports any potential conclusions. Integrating AI suggestions and assistance into existing processes requires attention to integration challenges, while also addressing the right amount of useful information that helps the human operators and researchers make quicker, confident decisions.
Federal agencies will be most successful with AI solutions if they have a holistic implementation strategy that includes examination of benefits, assessment of risks, attention to trust and transparency, and delivery of real outcomes. When implemented holistically and responsibly, AI technologies hold the potential to allow agencies to amplify and deliver more of what people do best, fundamentally reshaping how work is done and delivering clear benefits, quickly.
Accenture believes responsible AI can be addressed with a service design approach that grounds implementation with the core values of the agency’s policies and programs. This alignment between an agency’s commitment to the ethical delivery of its services, coupled with a transparent and accountable AI implementation, provides a consistent framework for maintaining the trustworthy nature of AI interactions with employees and citizens alike.
By transparently making sure AI reflects the best elements of human nature, agencies can address concerns up front and get the most out of their AI investments.
Incorporating a new technology that transforms the nature of public service and delivers innovative outcomes can be a challenge. But strong leadership can facilitate that process by establishing a strategic purpose around the use of AI, communicating its vision and support of AI programs, and managing the change to the organization as a whole—throughout the project life cycle and sustained by all stakeholders through go-live and beyond. It’s not so much about quick implementation; it’s about thoughtful, calibrated implementation, from beginning to end.
Developing strategies for AI that put people at the center and incorporate the human dimension is key to delivering real outcomes. Also key are taking into consideration an agency’s and the federal government’s policies, procedures, and service delivery for citizen-facing functions and committing to developing responsible AI systems that are aligned to ethical values that drive positive outcomes and empower people to do what they do best—imagine, create and innovate.
There is a real opportunity now for agency leaders to make government more efficient with AI, leading to better mission outcomes. We can leverage the immense power of computational technology to make experiences simpler and more natural for employees and citizens. With well-informed, augmented and AI-empowered decision makers, agencies will be able to advance government processes, establish a new culture for the workforce, and create a trustworthy environment with citizen-centered outcomes.
U.S. Department of Health and Human Services pilots Natural Language Processing
The U.S. Department of Health and Human Services has run a successful pilot to investigate the use of machine learning and Natural Language Processing (NLP) for increased efficiency in the processing of public comments on regulations. Agency staff and/or contractors currently use a manual and burdensome process that can be duplicative at times, especially when working under tight deadlines. For one sample Medicare & Medicaid Services rule, for example, it took more than 1,000 hours just to sort the public comments before the comments were even addressed. The test of a Content Analyst Analytical Technology (CAAT) tool that sorted comments demonstrated the potential to save thousands of employee hours, increase staff satisfaction and produce millions of dollars in savings. The project predicted an annual savings for one HHS agency of more than 300,000 employee hours when using the NLP capabilities.3
Accenture, “Technology Vision 2017: Survey Results US Public Services,” 2017.
Accenture, “Emerging technologies make their mark on public service,” 2017, https://www.accenture.com/us-en/insight-ps-emerging-technologies.
U.S. Department of Health & Human Services, HHS IDEA Lab/Office of the Chief Technology Officer, “Increasing Efficiency in Rule Making with Natural Language Processing,” http://www.hhs.gov/idealab/projects-item/increasing-efficiency-in-rule-making-with-natural-language-processing/.