The Business of Healthcare Podcast · Center for Healthcare Leadership and Management

Artificial intelligence in Healthcare

·36 min·2 clips

LineDance AI analyzes patient movement video to locate deficits, quantify them, and classify the problem.

1. The Business of Healthcare Podcast episode is about artificial intelligence in healthcare and how AI gets used, marketed, and misunderstood. 2. Bob Kaiser, Director of the Master's Program in Healthcare Leadership and Management at UT Dallas, hosts Mohamed El-Wazir, CEO of LineDance AI, because Mohamed builds AI tools for healthcare workflows. 3. The episode asks what AI is for in healthcare, and the core answer is that value matters more than hype or technical jargon. 4. Bob says AI is often used loosely and inaccurately, then introduces the phrase "AI washing" for older rule-based tools relabeled as AI. 5. Mohamed says the public has an incorrect perception of AI in healthcare because media coverage creates noise about what AI can and cannot do. 6. Mohamed compares AI's potential impact on healthcare to the discovery of penicillin, while saying the field is still in "the very, very early days." 7. Mohamed argues that healthcare professionals do not need to be technical to understand AI if they focus on value creation in a workflow. 8. He contrasts AI with BI, spreadsheets, word processors, and Microsoft Office to show how software tools have evolved over time. 9. He calls AI "superhuman" when it completes tasks faster or with fewer errors than a person using ordinary tools. 10. Mohamed describes LineDance AI's human movement comparative analysis, which uses patient movement videos to find deficits in how people move. 11. The system also quantifies how far a patient is from normal movement and classifies the problem as a motor control issue, a compensation problem, or something similar. 12. He says the tool supports physical therapy documentation so clinicians can spend less time on administrative work and more time with patients. 13. Bob summarizes the physical therapy use case as burnout work versus patient care, and Mohamed agrees that the documentation task is repetitive and time-consuming. 14. Mohamed says AI is best when it focuses on one clearly bounded task with a defined output, because that makes it easier to test. 15. The conversation turns to ChatGPT, and Mohamed says he first heard about GPT about three years ago when OpenAI was building it. 16. He explains that generative AI creates responses rather than retrieving prewritten answers, using summarizing a book as his example. 17. Mohamed says pre-training is part of the GPT name and describes the model as trained on large amounts of published text and internet content. 18. Prompt engineering is compared to online search, where specific wording changes results and overly broad prompts produce weak outputs. 19. Mohamed warns that fake images, fake voices, and deepfakes are growing risks, and he says new detection tools are appearing to identify them. 20. The tone is practical and explanatory, with Bob steering the interview through examples and Mohamed answering in plain language, and healthcare leaders who want concrete AI use cases would get the most from it while people looking for a deep technical discussion might skip it.

Listen to the show on

Apple Podcasts Spotify

Play clips from this episode More from The Business of Healthcare Podcast