OpenAI Says Scientists Are Increasingly Using ChatGPT as a Research Collaborator
A increasing role for artificial intelligence (AI) systems as collaborators in scientific research has been described by OpenAI, which contends that tools such as ChatGPT can aid researchers in solving challenging problems in a variety of fields, including biology, chemistry, physics, and mathematics. The 20-page report emphasizes how scientists are increasingly using AI models to help with data interpretation, literature synthesis, and experiment planning. It also underlines how improving AI’s research capabilities could speed up scientific and technical productivity and discoveries.
OpenAI Offers ChatGPT as a Research Partner
Millions of scientists, engineers, and mathematicians are already asking sophisticated questions and using AI to support scholarly work, from writing scientific text to debugging code and organizing experiments, according to anonymized usage data cited in the report, which was only shared with Axios. Opportunities for further integration of AI into the research process are suggested by OpenAI’s investigation, which shows that researchers mostly utilize AI for writing and communication duties, but less use it for rigorous analysis and computations.
According to OpenAI, average weekly message counts on complex science and mathematics-related topics increased by over 47% year over year (YoY) in anonymized ChatGPT interactions from January to December 2025. It was reported that the total number of mails increased from 5.7 million to about 8.4 million. The AI behemoth said in January 2026 that about 1.3 million users discussed complex science and math topics on a weekly basis.
“AI is being utilized more and more as a scientific partner, and its influence is growing in actual research environments. Advanced reasoning algorithms are being used by more researchers to solve open questions, decipher complicated data, and conduct experiments more quickly. Over the past year, that utilization has increased rapidly, and the effects are beginning to be seen in a variety of disciplines. Kevin Weil, VP of OpenAI for Science, stated, “We’re still early, but the rate of adoption and the caliber of the work suggest science is entering a new acceleration phase.”
Initiatives like OpenAI for Science, which aims to provide scientists and mathematicians with AI tools intended to speed up research processes from literature analysis to modeling and simulation, are consistent with the idea of AI as a collaborator. According to OpenAI, its objective is to create systems that naturally fit into scientific practice, assisting researchers in exploring concepts, testing hypotheses more quickly, and unlocking findings that would take years using conventional approaches.
