Three actions would greatly empower federal workers and executives to maximize the benefits of artificial intelligence.
Action 1: Emphasize training
Many federal workers worry that they are receiving inadequate training for AI. A 2019 Accenture survey, for example, found that 61% believed that they were poorly prepared to handle AI technologies. Insufficient funding is partly to blame—in the same survey, 55% of workers said that training costs were the biggest deterrent to upgrading their skills.
What kinds of skills will workers need to develop? Research by Accenture and others suggests that an aptitude for things like complex reasoning, creativity, and emotional intelligence will rank highly.
Narrow job descriptions will also become less prevalent, as federal workers are asked to do a wider range of tasks. Where more specialized skills are required, the emphasis will shift. Demand for engineers, data scientists, linguists, ethicists, and others who can train, sustain, and explain AI will rise.
A promising training model might be the kind of “pop-up” workshops used in 2019 by the Federal Cybersecurity Reskilling Academy, which made hands-on training in the field available to workers across the U.S. government. Innovative solutions to the challenge of reskilling can come from many places, including crowdsourcing and prize contests. The National Science Foundation used both techniques with its “Career Compass Challenge” (also launched in 2019), which invited the public to design worker-training programs. Federal agencies can also use online learning platforms, such as those available from Udacity and Accenture Academy, to continuously reskill their workforces in digital, IT, and analytics education.
To get the most out of AI, federal workers will need to be empowered to make the most of the opportunity. The same 2019 survey found that nearly three-quarters of workers received little guidance on how AI will affect their jobs in coming years. In addition to better top-down communication, workers’ input should be solicited on how AI might be used to make them more productive.
Action 2: Prioritize strategy
AI investment is rising fast, but budgets are not unlimited. Inevitably, the trendiest AI technologies won’t always be best suited to every agency’s needs. Agency executives should invest strategically.
To identify strategic priorities, better communication and collaboration between business and IT teams is needed. So is a long-term vision. To create one, seek input from workers on where they see potential AI application and collaboration opportunities. Focus on ways that AI might strengthen an agency’s mission. And think about how to connect AI with existing systems, to avoid duplication.
Another way to encourage strategic behavior is by making agencies less rigid. Federal agencies that can trim unnecessary bureaucracy, overcome departmental siloes, and reduce dependence on legacy systems are better able to shift priorities as conditions change. This flexibility can allow them to reconfigure operations to better align with AI’s strengths.
A 2018 Accenture survey of federal executives, for instance, found that nearly half believed that over-reliance on legacy systems hinder their ability to create innovative operating models and processes. The most agile federal agencies increasingly favor fluid open-source development, where platforms are virtualized, containerized, and automated. Such agencies encourage the prototyping of AI technologies, too.
Executives should also focus on scaling AI strategically. They might begin their efforts by answering this question: How could our teams, processes, and technologies be restructured to facilitate scaling?
Action 3: Rethink data
To get the most out of AI, agencies will also need to rethink how they collect, manage, and use their data. Begin by strengthening governance procedures and protocols, to better source and store the vast amounts of data that are required to train AI. Opportunities to converge data sources, across departments and agencies, can be identified as well.
Cloud-based computing—which offers big advantages on cost, efficiency, security, and flexibility—should become widespread. An upcoming Accenture study, Masters of Innovation: Lessons from Leading Public-Sector Organizations, however, found that only two-thirds of federal executives said that migrating their IT systems to the cloud in the coming years was a priority.