The artificial intelligence (AI) revolution is here. At the Department of Veterans Affairs, health care leaders are guiding its implementation in veterans health care with key ethical principles in mind.  

With VA being the nation’s largest integrated health system, AI is helping to treat more than 9 million veterans and showing opportunity for more, noted Assistant Undersecretary for Health for Discovery, Education and Affiliate Networks Dr. Carolyn Clancy. Clancy spoke at the agency’s AI summit in Washington, D.C. last week. 

Clancy emphasized the critical nature of the department’s adoption of AI and added that more than half of the VA’s 109 medical centers that support medical research also have an active study using some component of AI.  

“AI is relatively new, but putting technology to work for veterans and very importantly, for our employees, is really part of our institutional fabric,” Clancy said at the summit. 

Clancy outlined six key principles that the VA abides by when it comes to its usage of AI:  

AI must be purposeful, doing no harm and not sacrificing efficiency for accuracy and safety.  

AI must be effective and safe, with all systems designed and monitored to ensure accuracy and safety. 

AI must be secure and private, with built-in resiliencies against vulnerabilities and focused on ensuring privacy in data.  

AI must be fair and equitable. Clancy emphasized that because the VA has a diverse body of veterans, the response to providing them with health care must be diverse.  

AI must be transparent and explainable. 

AI must be accountable and rigorously monitored.  

Clancy highlighted some of AI’s applications in the VA health care system, including its partnership with the National Artificial Intelligence Institute (NAII) to help the VA’s National Surgery Program Office advance the capabilities of the VA Surgical Quality Improvement program.  

AI is being applied in scenarios to assist in developing phenotypic risk associations in cardiac and vascular surgery, building clinical decision support tools for chronic pain, prediction tools for cancer and progressive kidney disease and image-based assistant tools for pathology and radiation.  

Additionally, Clancy said AI is being used to better prevent, treat and diagnose cancer, as well as assisting in general diagnostics along with machine learning.  

While Clancy praised the VA’s AI usage efforts, she stated that tangible evidence will always be needed to prove that the technology is providing better outcomes for veterans. 

“Having a vision isn't enough. Even having proof of clinical effectiveness is not enough. We have to make sure that we are literally achieving the best outcomes for veterans,” Clancy said.