Author(s): Howard R Moskowitz*, Sharon Wingert, Sunaina Saharan and Stephen D Rappaport
The paper introduces a collaboration between generative AI and the researcher to understand how a medical professional can understand the thinking of a person newly diagnosed with a serious cancer. The approach uses ChatGPT 3.5 in a four-phase approach that can be done from start to finish at low cost, and in a repetitive fashion (iterations at each phase). In less than 12 hours, using the Mind Genomics platform, BimiLeap.com returns with insights and recommendations. The approach is set up so that each phase can be run repeatedly to build up a bank of deep knowledge. Each repeat effort for a phase is called an “iteration,” requiring 15-20 seconds of computer time. Phase 1 simulates a patient support group for patients with newly diagnosed cancers. Phase 2 simulates a series of AI-generated questions and answers regarding information that the patient would like to share with their personal physician, their closest friend, and a trusted cleric. Phase 3 shows a series of informative questions and answers which can educate the patient. Phase 4 shows a series of innovations recommended by AI, after AI has “digested” all the information it has provided to the user. Phases 1-3 are provided immediately after being created in an iteration. The innovations presented in Phase 4 are part of an extensive iteration-by-iteration analysis and expansion of the insights generated by AI for the specific iteration.
View PDF