Already integrated into many health care and public health applications, AI techniques can improve early detection of disease, help providers optimize treatment strategies and detect disease outbreaks.聽
Dean Mary Gallant of the Zuckerberg College of Health Sciences also sees how AI could potentially improve health equity鈥攁 fair opportunity for everyone to experience optimal health, regardless of race, ethnicity, gender, location or income.
鈥淏y processing and analyzing vast amounts of data at a speed that humans cannot, AI approaches can help us better understand underlying contributors to health outcomes and drivers of health disparities,鈥� says Gallant, a public health expert. 鈥淎I-driven diagnostic and treatment algorithms also have the potential to improve health care quality for all.鈥�
Gallant warns, however, that AI also has the potential to worsen health disparities and expand health inequities. For example, AI algorithms that are based on available data may lead to inherently biased results.
鈥淭he outcomes of AI are only as good as the underlying data it works with,鈥� says Gallant. 鈥淚f large datasets analyzed using AI don鈥檛 adequately include data from members of underserved populations or populations that experience health disparities, which they often don鈥檛, then the results of those data analyses will only benefit the groups represented in the data.鈥�
For example, if the data only includes people who are well-insured or who receive care at specialized health care organizations, then underserved groups won鈥檛 have access to improved outcomes, and disparities in health care could increase.聽
AI-driven technologies have already led to unfair outcomes for certain groups, says Gallant.聽
鈥淎s AI continues to be incorporated into health care and public health research and practice, it is essential that we keep considerations of health equity at the forefront of this work, so that the promise of AI can benefit all populations,鈥� says Gallant.