Signal OmniScientist: An Omni-Modal Omni-Discipline AI Scientist

Summary

Recent advances in foundation models have let AI scientists automate increasingly complete research workflows, spanning hypothesis generation, code execution and manuscript preparation. The authors argue, however, that covering more of the workflow alone does not provide access to the full evidence base that scientific discovery depends on. That is because existing systems typically reason over text, code, labels or precomputed summaries rather than raw multimodal evidence such as images, sensor data or experimental records. OmniScientist is proposed in response as an omni-modal, omni-discipline AI scientist system designed to reason directly over richer, discipline-spanning evidence types. The work positions itself as infrastructure for AI-driven scientific discovery generally, rather than a study focused on one scientific domain.