Many MindedBiologists now cite chatbots almost as often as specialized AI models
Epoch AI reports that the share of biology preprints acknowledging large language model use has grown exponentially since 2023. About 17% now do so, compared with about 19% that acknowledge specialized AI tools. The data was republished by HumanProgress on Nov 12, 2026.
The two kinds of tool have different trajectories. Acknowledgments of specialized tools, such as models for protein structure and design, image analysis, single-cell sequencing and genomics, have climbed steadily since 2021. AlphaFold, publicly available since July 2021, is the most commonly cited, appearing in roughly half of the preprints that acknowledge specialized AI. The LLM uses acknowledged include literature review, writing assistance and coding assistance. Acknowledgments of LLMs for research or analysis itself are still small: about 4% of biology preprints in September 2026, up from about 1% at the start of the year.
This matters for knowledge and health because analysis, coding and writing are bottlenecks that have historically favored well-staffed labs. A researcher without a bioinformatics team or a fluent command of English-language scientific prose can now get help with both. Specialized models like AlphaFold do something similar for structure prediction, a task that once demanded costly experiments. Cheaper analysis means more people can take part in the work that eventually leads to new treatments.
The caveats are real. These figures count acknowledgments in preprints, not measured use, so true usage may be higher or lower, since some authors may not disclose it. All numbers are approximate. The share using LLMs for actual analysis is still small, so watch whether that 4% keeps rising and whether it translates into results that hold up.
Source: HumanProgress