Co-Intelligent Design of Catalysis Research with Large Language Models: Hype or Reality?

Large language models are making a strong impact on broad fields of science and technology. Initial concerns like hallucinations have been largely mitigated in the latest models, enhancing their reliability. Co-intelligence, the synergy between human and artificial intelligences, has a strong potential in research when using large language models. This analysis focused on the question: Can large language models collaborate with chemists in the design of catalysis research? In a conversation where the author made a few prompts giving minimal guidance, GPT proved its ability to co-design a research project at different levels of detail. Besides a few, non-critical flaws, the project concept and workflows were provided by the model, making a major contribution to the research design. The assessment of the project by a committee including other models (DeepSeek and Gemini) was also investigated.