AI can enable powerful public sector innovations in diverse areas benefiting a wide range of citizens. Agencies should approach the technology with a holistic, value-driven strategy.
The City of Chicago is using algorithms to try to prevent crimes before they happen, while, in Pittsburgh, traffic lights using AI have helped cut traffic times by 25% and idling times by 40%. Meanwhile, some governments can use natural language processing (NLP) to monitor and analyze social media conversations to bolster counterterrorism efforts. Such examples illustrate how AI can improve government services by generating new insights and predictions, increasing speed and productivity, and creating new approaches to citizen interactions in areas as varied as national security, food safety, regulation, and health care. To fully realize AI’s benefits, government leaders can approach the technology strategically, aligning with the overarching organizational mission, drawing on lessons learned from past technology transformations, and incorporating important technical and managerial perspectives.
The vision for a holistic AI strategy should reflect clear alignment with broader agency goals, strategies, and policies. To articulate its value, focus first on its transformative capabilities. AI-based systems can make sense of large amounts of data quickly, which translates into reduced wait times, fewer errors, and faster emergency responses. It can also mine deeper insights, identifying underserved populations and creating better customer experiences. Using intelligent automation, AI can speed up and perform existing tasks, such as imagery analysis, beyond human ability. Employing predictive analytics, police can use NLP to spot otherwise-hidden crime trends and correlations in reports and records, for example, enabling quicker interventions. Given AI’s transformative potential, leaders can assure staff early on that AI will not eliminate their jobs; instead, they can demonstrate how AI-based tools such as language translation can enhance workers’ existing roles, freeing them from mechanical tasks in favor of more creative, problem-solving, people-facing work. For these professionals, as well as citizens and other stakeholders, this agency vision must also address the ethical considerations of AI and its value to the organization and its mission. Moving AI from pilot to production will involve new challenges, both technical and managerial, involving ongoing cleaning and maintenance of data; integration with a range of systems, platforms, and processes; employee training; and potentially operating model changes in parts of the organization. It is also critical to create a strategy to properly measure and track the value of proposed solutions—on deployment, for example, or usage—and to define performance standards in terms of accuracy, clarity, and transparency.
Capabilities to Execute AI Strategy
A government agency can have the right AI road map, technology, and funding and still fail at execution. The following actions can help build toward success:
Bridge the AI skills gap. Governments can acquire AI expertise in creative ways beyond recruiting, such as through upskilling and reskilling; creating AI guilds to share knowledge; hosting competitions and establishing partnerships to entice innovative solutions from outside the government; and forming partnerships with educational institutions and businesses to promote better understanding of AI and its benefits. Agencies can consider whether to develop AI talent internally, hire contractors, or use both. Relatedly, on the technical side, agencies must decide whether to create AI software in-house, purchase it off the shelf, or commission it as custom code.
Build a solid foundation. Appropriate architectures, infrastructures, data integration, and interoperability abilities are essential to a winning AI strategy. Agencies should test and modify infrastructure before rolling out solutions and determine whether their existing data center can manage the expected AI workload; often the answer is yes for a simple proof of concept, but no for a production solution. This also raises the issue of how data and cloud strategy may come into play. For example, the U.S. Department of Health and Human Services’ data strategy includes the consolidation of data repositories into a shareable environment accessible to all authorized users.
Establish Management Structures
Agency leaders can also develop management systems to validate specific AI initiatives, track performance, and scale projects from pilot to implementation. Some emerging tactics that can help bring AI strategies to fruition include:
Establish an AI operating model. Centers of excellence can develop best practices, provide training, and share resources and knowledge to identify and prioritize use cases and develop solutions. For example, the U. S. Department of Defense (DoD) established the Joint Artificial Intelligence Center with the goal of accelerating the delivery of AI-enabled capabilities, scaling AI departmentwide, and synchronizing DoD AI activities to expand Joint Force advantages.
Develop governance frameworks. Since data sharing is vital to achieving AI’s full impact, explicit governance models must address when and how data will be shared and protected. Government agencies increasingly use algorithms to make decisions—to assess the risk of crime, allocate energy resources, choose the right jobs for the unemployed, and determine whether a person is eligible for benefits, for instance. Well-defined AI governance structures explain how algorithms work and tackle issues of bias and discrimination.
Create deployment structures. A common trap in advancing a project from design through execution is “pilot purgatory,” in which projects fail to scale, often due to narrow designs, limited returns, or organizational resistance. One way to avoid such potholes is to focus initial AI resources on quick payoffs rather than more lengthy transformative projects. Further, before launching pilots, agencies should prioritize business issues and subject potential AI solutions to cost-benefit analyses.
Ensure data quality. Data is often stored in a variety of formats, in multiple data centers, and in duplicate copies. If federal information isn’t current, complete, consistent, and accurate, AI might make erroneous or biased decisions. All agencies should ensure their data is of high quality and their AI systems have been trained, tested, and refined.
Prepare to scale. Public-agency areas that generally call for significant AI-driven change include technical infrastructure, which requires high-bandwidth, low-latency, and flexible architectures to deploy AI applications; organization and team structures, with cross-functional teams making decisions based on fast-flowing, non-siloed data; talent management, with organizations training current workers, developing new retention strategies, and establishing new partnerships; and culture, with incentives for data-based experimentation and appeals to influential leaders to sponsor AI initiatives.