But, realizing the benefits of AI requires not only access to large-scale and comprehensive data to build reliable and actionable predictive models, but also public trust and engagement.
One of the ways to promote the effective application of AI is through partnerships between the public and private sector, drawing upon the scale and infrastructure from the public sector and the agility and technical talent from the private sector.
Motivated by a prioritization of AI throughout the federal government, the Department of Veterans Affairs (VA) has started using AI to improve the health and well-being of Veterans. Championing the largest integrated health care system in the country, the VA is uniquely suited to partner with the private sector to realize and scale the benefits of AI.
The VA is doing that through the implementation of AI Technology Sprints that bring the private sector, non-profits, and entrepreneurs together with the government to solve pressing problems through the application of AI and big data.
Cultivate Collaborative Partnerships
In November 2019, VA established the National Artificial Intelligence Institute (NAII) to cultivate collaborative partnerships in the private sector and higher education that lead to the successful piloting and eventual implementation of AI in government. While there are other parts of the federal government already working to advance AI, the NAII is unique in its access to large-scale healthcare data and the ability to coordinate across agencies with special authority to work with the private sector and research communities.
Promoting AI in large organizations can be challenging, especially in government because of the necessary role of oversight and the wide array of constituencies that are served. But one of the greatest strengths of large organizations is their ability to create visibility and connectivity among other actors to row in the same direction. The American AI Initiative and updated National AI R&D Strategic Plan developed by the White House have set the framework for furthering this area. Under recommendations from the Executive initiative, NAII successfully developed its award-winning AI-able ecosystem incentivization framework to help foster public-private collaborations and improve the health and wellbeing of Veterans.
One of the ways NAII encourages collaborative relationships is through AI Tech Sprint s. Centered around the application of AI research and development to VA data, tech sprints bring the private sector, non-profits, and entrepreneurs together with the public sector to help solve specific and timely problems.
Two of last year’s teams were recognized in a national competition, including winning first place in “creating the future of health” category. In 2019, the Girls Computing League (GCL), a non-profit organization committed to bringing technology to young women and students in the socioeconomic and racial minority, partnered with Amazon Web Services in the first AI Tech Sprint. They developed the Clinical Trial Selector (CTS), a web application that helps empower Veterans find clinical trials for which they may be eligible and benefit from.
Most clinical trial search engines require the patient to enter their medical conditions, which are then used to match them with trials. To simplify the process, GCL created an app that uses deep learning and natural language processing to allow Veterans to securely login to their accounts and directly search for trials using the information already contained in their records with no extra steps. In other words, the app automates an otherwise time-intensive and complicated process, which helps not only the patient, but also public health interventions at large by facilitating the matching of supply and demand.
NAII’s AI Tech Sprint Handbook provides details and explains the tech sprint process and value proposition.
Looking Forward—Helping Unreached Veterans
Although the VA serves over 9 million Veterans, there are millions more who may benefit now or in the future from VA services. And yet, learning how to serve these Veterans is challenging because it is so difficult to obtain data on them. Theories from behavioral economics suggest that outcomes that are hard to measure are likely to get pushed to the side, making it all the more important to use technology to improve our tracking and delivery of care to those and their families who gave it all.
Consider data from the Census Bureau’s American Community Survey (ACS) that shows the number and percent share (of the population) of Veterans between 2005 to 2017. Although the number of Veterans is declining, there is a gap between the roughly 18 million in total and the 9 million engaged in the VA. And, since the number of Veterans has been declining over time, the pattern highlights the especially precarious position they find themselves in. If the federal government doesn’t leverage technology to improve the design and delivery of Veterans services, outcomes could deteriorate.
Motivated by these challenges and the prospective benefits of AI, the new 2020 AI Tech Sprint is focused on designing interventions aimed at engaging these Veterans who are not currently served by VA. The AI Tech Sprint will be a 12-week engagement in collaboration with industry and academic partners to develop and deploy AI tools that leverage federal and other VA-generated data. Those interested in participating can apply here. This year, the VA will be hosting a competition that will award up to $100,000 in prizes. These AI Tech Sprints also provide opportunities for future partnership too.
The tech sprints leverage insights from behavioral economics and psychology that emphasize the importance of partitioning work into pieces and bringing participants together with different experiences and resources to the same table. Moreover, because of the fast-paced nature of the sprint, tech sprints provide an opportunity to iteratively pilot new technology and receive rapid feedback to either accelerate the development or pivot.
Although the AI Tech Sprint is applied with a focus on Veterans, the underlying process is scalable and applicable across government. There are many tools available for promoting innovation, but the AI Tech Sprint looks particularly attractive for bringing people together around a specific goal and timetable.