A new study released today by MeriTalk finds that Federal agencies are highly aware of challenges they face in harnessing and analyzing data, and that artificial intelligence (AI) offers an opportunity to change how government handles and processes data. And the desire of agencies to more fully embrace AI seems to be making notable progress in clearing away practical obstacles to its increased use.
Data-Driven Government? Not There Yet
Federal agencies know the importance of harnessing data. Almost all Federal IT managers, 95 percent, either have or are working on a comprehensive data strategy, and 87 percent say data is key to delivering better outcomes, the study finds.
However, only 17 percent of those surveyed think their agency is “completely successful” at using that data to gain insights into their organization. Data management challenges are familiar to many agencies, and they can hamper efforts to become more data-driven.
“Agencies need to begin by developing a data-centric strategy, and then work their hypothesis with machine learning, and they will get the artificial intelligence gains they are looking for,” said Gary Newgaard, vice president, public sector for Pure Storage, which underwrote the study. “The basic foundation is having a data structure set up so you can access a variety of data sets.” Newgaard gave an example of uploading pathology slides from past decades to gain insights into disease. “If you have hundreds of thousands of cancerous scans that are in a different format, that is a lot of historical data and lessons learned that you can take advantage of.”
Agencies feel held back from utilizing their data by a range of issues including security challenges, speed of access, quality and consistency issues-which particularly concerned C-level leadership-and the cost and availability of storage. Ultimately, 73 percent see their challenge as not just getting the data, but analyzing and interpreting it to turn it into actionable information.
AI on the Horizon
However, AI offers a “superhighway” to get past these roadblocks. Seventy-seven percent of those surveyed believe that AI will be a game changer for how government handles data, and 61 percent say they have a problem that AI could solve. As one respondent put it, “we are drowning in data, and AI should be our lifeboat.” Looking 10 years down the road, respondents want to see AI used to manage services efficiently, identify cyberattacks, and provide recommendations for human employees.
Newgaard noted the potential use cases of monitoring traffic, aggregating healthcare data, and analyzing vast amounts of threat intelligence data. “Time to answer is really the benefit of artificial intelligence, along with being able to shorten your decision-making cycle.”
The excitement for AI goes beyond talk and into planning, as 48 percent of Federal IT managers and 69 percent of C-level leadership say AI is a part of their agency’s technology roadmap. Preparation today pays off down the road, as agencies with a comprehensive data strategy in place are three times more likely to include AI in their agency’s future plans.
IT leaders in the Department of Defense (DoD) are also more likely to be preparing for AI, with 83 percent of DoD respondents saying they are taking action on AI. DoD agencies are also further down the road, with 56 percent of agencies in the planning process, compared to 29 percent of civilian agencies. The DoD continues lead the Federal government on AI adoption with initiatives such as the Joint Artificial Intelligence Center.
However, Newgaard noted that the DoD’s advances can offer lessons for civilian agencies.
“If you look at the DoD, most AI work is being done around intelligence. But if you take a closer look, they have broad responsibility for medical support and healthcare and are doing great work around genomics research – those lessons learned can apply to NIH or CDC. There’s lots of cross-pollination that can be done.”
Emerging Technology and Knowledge Lead the Way
Agencies won’t be able to take full advantage of AI overnight. The study also highlights the growing adoption of emerging technologies such as big data analytics, automation of data ingest and transformation, and deep learning. Agencies that are using these technologies are motivated by some of the same end goals as AI; 50 percent are adopting these technologies for operational efficiencies, 39 percent are looking towards cybersecurity, and 35 percent use them to help strategic decision making. Eighty percent of Federal IT leaders say that success in these efforts will lead to more robust AI adoption.
“What we’re seeing across most of government is, once you’ve identified your data structure and the data sets and they’re available, you have to start small. You can’t just jump into the deep end of the pool,” said Newgaard.
On the journey to AI, agencies expect to encounter challenges such as scarce funding, a shortage of Federal guidance and a lack of understanding about potential benefits. However, the benefits are worth the struggle. Early adopters of AI in government note that they already have fewer challenges with data management, security, and bottlenecks. Those who have already implemented AI pilot programs advise agencies to start small, understand the equipment costs, and get their team on board.