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

05 June 2025, Version 2
This content is a preprint and has not undergone peer review at the time of posting.

Abstract

Large language models are making a strong impact on broad fields of science, pushing the efforts to minimize hallucinations and to support specialized language. Co-intelligence, the synergy between human and artificial intelligences, can have a strong potential in research through the use of large language models. This Viewpoint focused on the question: Can large language models collaborate with chemists in the design of catalysis research? In a conversation based on a few prompts giving minimal guidance, ChatGPT-4o proved its co-design abilities at different levels of detail, making a major contribution to the definition of a research project concept and its workflows. Relevant flaws were also observed, highlighting the importance of involving critical thinking from highly-educated researchers. The assessment of the results by ChatGPT-4o itself and other large language models, including DeepSeek and Gemini, was also explored.

Keywords

Co-Intelligence
Large Language Models
Human-in-the-Loop
Research Design

Supplementary materials

Title
Description
Actions
Title
Supporting Information
Description
Transcripts of conversations with large language models (ChatGPT, DeepSeek, and Gemini).
Actions

Comments

Comments are not moderated before they are posted, but they can be removed by the site moderators if they are found to be in contravention of our Commenting Policy [opens in a new tab] - please read this policy before you post. Comments should be used for scholarly discussion of the content in question. You can find more information about how to use the commenting feature here [opens in a new tab] .
This site is protected by reCAPTCHA and the Google Privacy Policy [opens in a new tab] and Terms of Service [opens in a new tab] apply.