Geographic imbalances could hinder progress of AI applications in life sciences

Asia and Northern America dominate in the global production of Artificial Intelligence (AI) research in the life sciences, finds new research from Mannheim Business School, with the US and China producing 45% of all the research from 2000 to 2022.

Life sciences include biology and sub-disciplines, vital in health and medicine applications. AI could transform medicine through life science research, but the extent and pace of this transformation also depends on geographic concentration.

Professor Marc Lerchenmüller and Dr. Leo Schmallenbach from Mannheim Business School, alongside Professor Till Bärnighausen from Heidelberg University, analysed over 390,000 AI life science research publications from 2000 to 2022 and 14.5million associated citations, focused on six world regions.

Although Asia and Northern America lead in number of total publications, Europe and Northern America produce most of the research appearing in high-ranking outlets, contributing 70% of the total research there. High-ranking outlets include the top three journals within a given research field or top tier conference proceedings publications.

Using the number of forward citations received as an indicator of influence on subsequent research, Northern America, Europe, and Oceania were found to produce research most relevant to advancing AI in life sciences.

Latin America and Africa account for less than 5% of global research. This is despite these two regions being home to more than 25% of the world population.

“If research remains geographically concentrated, data evolves in an unbalanced fashion. This imbalance could lead to biased AI models producing biased recommendations,” says Prof. Lerchenmüller. “Patient populations have incredibly diverse characteristics. To mitigate the risk of AI informed medical care being biased towards certain demographics, straddling these different characteristics requires more and accelerated research and training data sets globally.”

International collaborations were also found to produce more relevant research than national collaborations. AI life science research produced through international collaborations receive 20% more citations than research from national collaborations. “Despite this, the rate at which scientists collaborate internationally is declining,” says Dr. Schmallenbach.

Greater integration of global expertise and international collaboration should be promoted to help utilise the benefits of AI research and contribute to improving global health.

These findings were first published in the journal Nature Communications.

/ENDS

For more information, a copy of the research paper, or to speak with Dr. Schmallenbach, please contact Kyle Grizzell from BlueSky Education at +44 (0) 1582 790709 or kyle@bluesky-pr.com