As AI tools continue to move into healthcare and the workplace, HSI is supporting new research that takes a more grounded view of how these technologies affect workers, organizations, and patient outcomes.
Two newly funded projects, selected through the HSI Research Fund’s most recent call for proposals, focus on the real-world implications of AI adoption. These projects join a broader portfolio of 29 studies led by 16 researchers.
One project, led by Jackson Lu, examines whether generative AI can provide meaningful psychological support to workers. While AI-based mental health tools are increasingly common, there is limited rigorous evidence on their effectiveness in workplace settings. This study will conduct a longitudinal field experiment comparing voice-based AI counseling, professional human counseling, and a non-counseling control group in a high-stress work environment. The research aims to identify where AI support may be effective, where it falls short, and what tradeoffs employers should consider when expanding access to mental health resources.
A second project, led by Erin Kelly, Kate Kellogg, and Zana Buçinca, focuses on how AI is implemented within healthcare systems, particularly among frontline and lower-wage workers. Partnering with a large health system and a labor union, the team will study recent AI and analytics deployments using case studies, focus groups, and quantitative analysis. The goal is to understand why some tools are successfully adopted while others stall, and what meaningful worker involvement in technology decisions looks like in practice.
These new efforts complement ongoing HSI-supported research led by Joseph Doyle, which examines whether health information technology improves patient outcomes. Focusing on Health Information Exchange (HIE), the study uses state-level policy variation to estimate causal effects of adoption. The findings to date indicate that HIE implementation is associated with an 18% reduction in mortality from flu and pneumonia and a 4–5% decrease in hospital readmissions. As adoption expanded, these improvements are estimated to have contributed to approximately 27,000 lives saved annually, likely driven by better care coordination and more effective public health response.
Taken together, the newly funded projects reflect HSI’s focus on generating practical evidence about how AI affects workers and organizations in real-world settings. At the same time, ongoing research on health information technology continues to demonstrate the measurable impact that data infrastructure and care coordination can have on patient outcomes.