From Workflows to Teamflows in eScience: Integrating Collaborative Science, AI, and Computing for Impact at Scale

eScience is now an example field for conducting science using advanced computational methodologies, middle-ware and workflows. With the ever-increasing data intensity in various scientific disciplines and the opportunities provided by novel artificial intelligence (AI) and computing techniques, there is a pressing need for eScience to adapt and develop holistic approaches for AI-integrated science and interdisciplinary collaboration between science, AI and computing experts. This paper discusses the transformative shift needed in the conduct and tools of eScience from traditional workflows, often linear, pre-defined to coordinate task execution and individually developed, to integrated “teamflows”, a term coined to encapsulate methods for effective team collaboration and communication, dynamic data and service integration, and use-inspired thinking. In the AI in science era of scientific investigation, this “future of eScience” special track paper: 1) emphasizes the vital importance of eScience to the success of collaborative AI in science teams; 2) recites some requirements for open data and services for AI-enhanced workflows, team science and incentive structures to create a more inclusive and efficient discovery environment; and 3) discusses impact as a measure of success in eScience.