Personalized Auto-Research: Towards a True AI Co-Scientist

Summary

AI co-scientists are beginning to transform research by generating hypotheses, retrieving related work, designing experiments, executing code, and drafting papers, but current systems ignore the individual researcher, optimizing generic metrics like novelty and reviewer scores. This paper introduces personalized auto-research, which conditions every stage on a representation of the researcher, arguing that personalization is essential for a genuine co-scientist. The proposed framework features three components: graph-grounded researcher representations, full-pipeline personalization, and individualized evaluation. It highlights the failure mode where different researchers receive identical output, erasing tacit knowledge crucial for novel ideas, and discusses open challenges.