Using AI in Health Care to Enhance Human Abilities

Fourteen hours. That’s how much time per week the average physician in the U.S. spends on prior authorization requests, according to the American Medical Association. Sometimes, the clinical evidence amounts to hundreds of pages, which the insurance provider must read through and interpret before making a decision. It’s an example of a burdensome and time-consuming process where, inevitably, the humans involved will still miss something. 

“This is where artificial intelligence can excel,” says Andreea Bodnari ’10. The founder and CEO of ALIGNMT AI, a start-up that provides healthcare companies with automated AI risk monitoring and compliance, Bodnari has devoted most of her career to building AI systems that augment human capabilities. 

“Our brains excel at specific tasks, but not all. Analyzing massive datasets to draw conclusions is not our forte,” she says. In the healthcare space, AI augmentations are now used to help with this type of clinical reasoning, as these tools can quickly summarize large amounts of data and help providers with decision making, particularly in areas where humans are known to have psychological biases.

For Bodnari, who was born in Romania, researching AI techniques that boost human efficiency and accuracy is nothing new. Her fascination goes back to her undergraduate days at WPI.

“My time at WPI was instrumental in shaping my path,” Bodnari says. 

As a computer science major, she started conducting research and developing practical applications of artificial intelligence in healthcare. When she worked with UMass Chan Medical School on an undergraduate research project, she got the opportunity to contribute to the design of clinical decision support tools that were used to analyze cancer patient profiles and predict the outcomes of certain procedures for those patients. 

“The relevance and impact of our published research solidified my passion and prepared me for the rigors of a top PhD program, ultimately leading to my acceptance at MIT,” she says.  

That’s where her “serendipitous” foray into natural language processing (NLP) took place. Although she’d wanted to focus more on proteomics and clinical biomarkers, the available research funds at MIT were for digital biomarkers and clinical NLP. So, she went in that direction. “The field was still emerging, which made it incredibly interesting,” she says. In 2014, she graduated with a PhD in machine learning from MIT’s Computer Science and Artificial Intelligence Lab.

Bodnari went on to lead healthcare AI products at Google Cloud and UnitedHealth Group. From those experiences, she learned almost as much about consumer education as she did about new AI systems. 

“My time at Google Cloud highlighted the importance of translating cutting-edge technology into business solutions and actively educating the market,” she says. The big lesson was that companies must teach consumers how to apply any new technology in their own lives or businesses. Otherwise, adoption will be slow.

She cites the “Intel Inside” campaign of the early 1990s, which educated consumers about the invisible processors that power computers, as a “masterstroke in democratizing technology.” 

“Intel transformed the PC from a technical marvel into a household essential,” she says. “As we stand on the precipice of the AI revolution, a similar approach that resonates with the average consumer is imperative. By making technologies like generative AI accessible, understandable, and beneficial in everyday life, we can ignite a new era of innovation and transformation.”

A key part of educating consumers, she discovered, is building trust. In her experience, some AI solutions have been left on a shelf not only because people don’t understand their use cases, but also because they don’t trust their performance or outcomes yet. But that kind of trust requires transparency, and establishing transparency in AI is a complex challenge. 

Tackling that challenge is what Bodnari is focused on today. Her company, ALIGNMT AI, is building the bedrock for AI transparency, automating compliance and risk mitigation requirements for AI applications in healthcare. The platform’s clear measurement standards and continuous performance reporting will be vital in defining “good performance” for AI in healthcare settings going forward.

Bodnari is currently a member of the WPI Executive Advisory Board for Data Science and Artificial Intelligence, where she helps develop strategic research and education initiatives, including the university’s new Master of Science in Artificial Intelligence program. As AI becomes increasingly interdisciplinary, she believes that new dimensions such as ethics, human-machine collaboration, and regulatory affairs deserve greater emphasis in the curriculum. In particular, she advises that exploring how to integrate ethical principles into the very design of AI systems should be an essential part of training the next generation of AI professionals. 

But when it comes to advising the general population on how to navigate the future of AI, she suggests a different strategy: Read books.

“What sets us apart as humans is our ability to reason and to have a different type of intelligence that is not based on large amounts of data. That type of intelligence is shaped through abstract thinking and conceptual thinking, which are the muscles you train while reading,” she says. “That’s how we will continue to advance our thinking and imagination and identify new ways to shape this technology for society’s benefit.”