The Department of Veterans Affairs (VA) is in the process of developing and piloting several projects in artificial intelligence (AI) as the agency prepares for the next frontier in health IT – human-machine interactions – the agency’s AI lead said on July 12.
“There are so many new opportunities in the field of AI and these new developments have so many potential benefits,” Gil Alterovitz, the VA’s AI director, said during a virtual summit organized by ATARC.
The guiding mechanism in the development of these new AI projects is the agency’s Trustworthy AI Checklist. The checklist, according to Alterovitz, was developed through feedback received from previous surveys on what makes up trustworthy AI.
The surveys – conducted by the VA – were taken by leading AI experts in the Federal, academic, and private sector.
“This AI checklist is essentially an educational tool that we can look to as we begin to develop more uses for AI, ensuring we implement AI in an ethical and secure method. So far, we have been able to implement secure and ethical measures from the beginning, rather than at the end when we already had an AI model built and implemented,” Alterovitz said.
Alterovitz highlighted that progress on the human-machine frontier depends on the Federal, academic, and private sectors continuing to interact and share their insight on different AI uses, and what makes secure and trustworthy AI.