Disclosures: The authors have disclosed no relevant conflicts of interest. Objectives: Training in pathology involves learning morphology of several cell types utilizing images and categorizing them to specific patterns. In general, since there are immense subjective differences in the interpretation of images, there are also severe subjectivity in teaching specific images to residents. While machine learning is currently intensely developed, we hypothesize that machine based teaching will remove subjectivity in learning and thus propose to study machine teaching skills. Methods: We use machine based image analysis and counting of bone marrow cells in this study. Two different artificial intelligence platforms were utilized. In the Scopio platform for bone marrow differential counting, machine based automated classification will be reviewed by an expert; cells are re-categorized. Subsequently, trainees are 'taught' by the Machine and by an iterative training process in stage I: Expert-> Machine->Trainee and subsequently implementation stage II: Machine->Trainee with Expert oversight. In the second platform, Deepathology Studio, identification of Bone marrow megakaryocytes by learner will be used as a qualitative skill. The expert teaches the machines for identification of megakaryocytes in core biopsies. Subsequently, the machine identified megakaryocytes will be utilized to learn bone marrow megakaryocytic morphology by the learner (trainee). Results: Image based identification of bone marrow by Scopio and categorization was utilized by the machine to teach morphology of 9 different myeloid cells, lymphocytes, plasma cells and 2 stages of erythropoietic cells. Core biopsy megakaryocytes can be routinely identified by machine utilizing the Deepathology Studio and can be used to 'machine teach' for training. Conclusions: While early in development, machine teaching can be akin to game based learning and trainees like the electronic format of learning. In addition, the objectivity increases. However, the major limitations are false positive categorization by machine and attrition effect will enhance wrong learning. Thus, close supervision by expert is essential during the early training phase with routine monitoring till knowledge acquisition.

MT-HL: Machine Teaching Human Learning - Improving Consistency in Image-based Pathology Teaching