Analysis of approximately 75 million publications reveals that those using AI are more likely to be “hit journals.”

From designing new drug candidates in medicine to developing new tax policies in the social sciences, the benefits of artificial intelligence (AI) in scientific research are omnipresent.

Just this week, two scientists known for their pioneering research in AI won the Nobel Prize in Physics, and three scientists won the Nobel Prize in Chemistry, which recognizes the use of advanced technologies, including AI, to predict the shape of proteins. However, despite its rapid progress and broad applications, many researchers do not systematically understand how AI can benefit their research, and skepticism remains about its ability to advance science across fields.

A new study from Northwestern University analyzing 74.6 million publications, 7.1 million patents and 4.2 million university course syllabi finds that articles that use AI have a “premium of impact on quotes.” However, the benefits of AI do not extend equitably to women and minority researchers, and to the extent that AI plays a larger role in accelerating science, it may exacerbate existing disparities in science, with implications for building a diverse, equitable, and inclusive research workforce.

The research team, led by Dashun Wang and Jian Gao of the Kellogg School of Management, developed a measurement framework to estimate the direct use and potential benefits of AI in scientific research by applying processing techniques natural language (NLP) to these large datasets.

Wang is professor of management and organizations at Kellogg and industrial engineering and management sciences at McCormick, director of Kellogg’s Center for Science, Science, and Innovation (CSSI), and co-director of Kellogg’s Ryan Institute on Complexity . Gao is a research assistant professor at Kellogg CSSI.

The study, titled “Quantifying the Use and Potential Benefits of Artificial Intelligence in Scientific Research,” was published Oct. 11 in the journal Human behavior.

“These advances raise the possibility that, as AI continues to improve in precision, robustness and scope, it could bring even more significant benefits to science, propelling scientific progress across a broad spectrum of research areas while significantly increasing the innovation capabilities of researchers,” Gao said. said.

Most notable research

The study reveals that recent AI successes, across the board, have been remarkable for research. There has been an increasing use of AI in disciplinary research since 2015, as evidenced by the mention of AI-related terms (such as “artificial intelligence”, “deep learning”, and “convolutional neural network”) in the title or abstract of publications.

From 2015 to 2019, disciplines including computer science (37%), engineering (24%), physics (24%), biology (22%), psychology (24%), economics (14 %), sociology (30%), and political science (27%) all showed a large increase in direct use of AI scores due to the development of new AI capabilities.

Researchers look at the number of times an article is cited and define a “hit article” as being in the top 5% of citations to articles published in the same field in the same year. Regardless of discipline, disciplinary articles that mention AI-related terms in their title or abstract receive more citations, being more likely to be a success, and receive a higher fraction of citations from others disciplines.

“In addition to its expansion, the use and benefits of AI in research are pervasive across disciplines, but we have seen systemic misalignment in AI education,” Gao said. “Investment in AI in higher education is not keeping pace with the benefits of AI in science.”

These findings suggest that the supply of AI talent and knowledge across most disciplines appears insufficient given the benefits these disciplines can derive from AI capabilities, highlighting a significant gap between the use of AI and AI training.

“The use of AI in scientific disciplines has advanced across science, while the focus on AI education to upskill future scientists in each discipline has lagged,” Gao said.

Underrepresented groups in STEM fields

The study also highlights the unequal effects on women and minority researchers that could result from the continued increase in the use of AI in scientific research.

“Historically, we know that women and minorities are less represented in certain fields, particularly in STEM fields,” Gao said. “We found that as the use of AI in science continues to grow, these same groups are less likely to benefit from new technologies.”

The researchers suggest that investing in ensuring training behind AI is equitable could have a positive impact on closing the demographic gap.

What’s next?

As AI rapidly evolves, researchers say we need to continually monitor and update its benefits for science.

“Women and minorities benefit the least, so how can we alleviate these demographic disparities? Gao said.

The research team’s analysis supports the hypothesis that collaboration between domain experts and AI researchers could represent a significant way to facilitate the use of AI in science and bridge the gap between the use of AI and AI training.

“There is benefit to increasing AI training across disciplines, which would likely help disciplines develop domain-specific AI expertise, allowing them to benefit more and more quickly from advances in AI. AI,” Gao said.