More than 13 percent of biomedical research abstracts submitted globally in 2024 show evidence of having been drafted with the assistance of artificial intelligence, according to a study released Thursday.
The surge in AI-assisted submissions is creating challenges for peer review and raising concerns about the quality of scientific literature, researchers say. The findings come as generative AI tools become increasingly sophisticated and accessible to researchers worldwide.
The study, released February 15, 2026, does not identify the specific AI tools used to generate the abstracts, but notes the increasing capabilities of such technology. The proliferation of AI-generated content is prompting discussion about how to maintain the integrity of the scientific process.
Japan and the United States are collaborating to address these challenges, with a framework established in April 2024 to leverage generative AI for scientific advancement. The initiative involves deploying Japan’s Fugaku supercomputer alongside the Aurora supercomputer in the U.S., according to a joint statement from both governments. The collaboration aims to modernize research practices, but similarly implicitly acknowledges the need to manage the risks associated with AI-generated content.
Satoshi Kawahara, a staff writer for the Nikkei, reported on both the surge in AI-assisted papers and the U.S.-Japan collaboration. Kawahara also reported in July 2023 on Japan’s efforts to emulate SpaceX’s iterative development model in its next-generation rocket program, suggesting a broader national focus on rapid innovation and adaptation.
The increasing reliance on AI in scientific writing is not limited to biomedical research. The potential for AI to accelerate drug development is also being explored, with Fujitsu utilizing generative AI to expedite the process, as reported by Nikkei in April 2024.
While the U.S. And Japan are actively exploring the benefits of AI in research, concerns about quality control remain. No immediate solutions have been publicly announced to address the potential for substandard research literature resulting from AI-assisted submissions.