This conference brought experts in areas of STEM domain-specific learning, assessments and measurement, learning technologies, computer sciences, and pedagogy to discuss and address the challenges of applying AI in revolutionizing STEM assessments and identify future research directions. It generated knowledge of integrating AI-based STEM assessment practices with domain-specific learning theories, validity theories and assessment design principles, technology integration theories, and pedagogical theories focusing on assessment practices. By working with globally well-known experts in assessments, AI, and STEM education, this conference tackled the critical challenges in four identified areas. The participants shared empirical findings, as well as challenges identified in the research, with a broad audience. By collaborating together, the conference identified the potential solutions to the challenges and future research directions to increase the broad use of AI-based assessments in STEM education. The proceeding book, Uses of Artificial Intelligence in STEM Education, will be published by Oxford University Press.