Agata Ciesielski, a fellow in the White House Presidential Innovation Fellowship (PIF) Program, worked with the National Artificial Intelligence Institute (NAII) at VA to help develop the VA Strategy for Artificial Intelligence (AI)† She also co-authored a white paper, “AI strategies and synergy in the federal space”, with NAII Director Gil Alterovitz, and contributed to implementation guidance for Executive Order 13960 on Trustworthy AI†
However, her interest in AI was “a mistake,” Ciesielski says. During her undergraduate fellowship with the National Aeronautics and Space Administration (NASA), the principal investigator assigned her to work on neural networks, thinking it was close to her major in neural engineering. As she learned more about neural networks, she became fascinated with AI’s mix of math, statistics, and psychology.
Ciesielski collaborated with colleagues at NAII because she had previously worked on AI applications in healthcare. As the nation’s largest integrated healthcare system, VA provided countless opportunities to advance AI in healthcare. VA has several big data repositories, including the largest genomic knowledge base in the world linked to healthcare information that can be used for AI. “This positions VA as a high-impact place for a population of people who have given so much to protect our country,” Ciesielski said.
An infrastructure for AI solutions
She notes that there is an urgent need to create the infrastructure to amplify AI solutions. For example, if an algorithm shows promise during testing, there must be a system to try it on larger populations and perform system updates. She adds: “VA is well placed to develop these complex systems. [With such a large system,] VA has potential medical diagnostic solutions, a large amount of data to create the solutions, and the basic infrastructure to deploy enterprise-level diagnostic tools.”
Ciesielski explains that as AI becomes more integrated in the public sector, there is an “urgent need for a common language and approaches”. While there have been successful AI applications and demonstrations at the project and program level in the federal government, efforts are often siled. Organizational level standards are needed to improve solutions. Greater coordination between agencies would help reduce duplication and maximize impact.
Ciesielski co-developed the White Paper on AI Strategies in the Federal Space to help agencies understand the government’s holistic approach to AI, while developing their own AI strategies. “The research provides a framework and each agency should tailor the AI approach to its organizational mission,” says Ciesielski. The document has been shared with the White House, the Chief Information Officer/Chief Data Officer Councils and other government policy-making groups.
`The beauty of the program’
When asked what she would tell potential fellows, Ciesielski replied: “The beauty of the program is that PIFs are placed and challenged to work on major inter-agency issues that we as individuals are passionate about. PIFs connect with other fellows and federally employed PIF alumni like Dr. Gil Alterovitz, who empower fellows, we are introduced to leaders to help make great things happen.”
“Working with the NAII was beyond fantastic,” says Ciesielski. “Not only is there the opportunity to work on major impact issues, such as improving AI solutions across the VA network, but Gil’s involvement and strong connections with inter-agency efforts ensure all work is in sync with the federal, industrial and academic cutting-edge efforts and practices.”
Applications to become a Presidential Innovation Fellow for 2023 are now open until July 8 at presidentialinnovationfellows.gov/apply