The 'Med AI' Capsule Podcast by Dr Avneesh Khare

In this episode of The Med AI Capsule, we explore how AI in healthcare is evolving from passive chatbots to autonomous agents that can execute entire clinical workflows. We break down how these systems work, where they add value in reducing cognitive load, and why clinicians must shift from evaluating outputs to auditing workflows—while maintaining strict oversight to prevent hidden, compounding errors.
Note: This audio podcast is an extension of The Med AI Capsule newsletter by Dr Avneesh Khare. You can explore more at www.avneeshkhare.com.
Disclaimer: This audio podcast was generated using AI tools and is based on content from the related newsletter issue. While all the information is curated from reliable sources, it is only intended for educational and informational purposes. AI can make mistakes – please check all the important information at your end. For any issues, please reach out at [email protected].


This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit avneeshkhare.substack.com

In this episode of The Med AI Capsule, we unpack multimodal AI through a simple walkthrough of how modern AI systems can integrate multiple types of clinical data—imaging, labs, EHR records, signals, and clinical notes—to build a more complete understanding of a patient. We explore why traditional single-data AI systems hit a “context ceiling", how multimodal models translate diverse inputs into a common representation and fuse them to uncover hidden clinical patterns, and what this means for improving prediction and decision support across specialities. We also touch on the real-world considerations—including explainability, bias, data integration challenges, and privacy—while emphasizing the central idea: multimodal AI is best seen as a clinical copilot designed to augment, not replace, physician judgement.
Note: This audio podcast is an extension of The Med AI Capsule newsletter by Dr Avneesh Khare. You can explore more at www.avneeshkhare.com.
Disclaimer: This audio podcast was generated using AI tools and is based on content from the related newsletter issue. While all the information is curated from reliable sources, it is only intended for educational and informational purposes. AI can make mistakes – please check all the important information at your end. For any issues, please reach out at [email protected].


This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit avneeshkhare.substack.com

In this episode of The Med AI Capsule, we break down federated learning through a simple, conversational walkthrough of how hospitals can collaboratively train AI models without ever sharing patient data. The discussion covers why this matters for privacy, how the workflow actually operates in practice, the safeguards needed to prevent information leakage, and the real-world limitations that still require clinical oversight. It’s a clear, practical primer on the idea of collaboration without data sharing—and why that matters for building trustworthy medical AI.
Note: This audio podcast is an extension of The Med AI Capsule newsletter by Dr Avneesh Khare. You can explore more at www.avneeshkhare.com.
Disclaimer: This audio podcast was generated using AI tools and is based on content from the related newsletter issue. While all the information is curated from reliable sources, it is only intended for educational and informational purposes. AI LLMs can make mistakes - please check all the important information at your end. For any issues, please reach out at [email protected].


This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit avneeshkhare.substack.com

In this episode of The Med AI Capsule, we decode RAG—how it connects AI to evidence. Instead of relying on training data alone, RAG retrieves current clinical guidelines and research before answering. We explore why this matters for clinical reliability and what still requires human judgment.
Note: This audio podcast is an extension of The Med AI Capsule newsletter by Dr Avneesh Khare. You can explore more at www.avneeshkhare.com.
Disclaimer: This audio podcast was generated using AI tools and is based on content from the related newsletter issue. While all the information is curated from reliable sources, it is only intended for educational and informational purposes. AI LLMs can make mistakes - please check all the important information at your end. For any issues, please reach out at [email protected].


This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit avneeshkhare.substack.com

In this episode of The Med AI Capsule, we look at how medicine is racing ahead and why explainable AI has become the central requirement for safe adoption. We unpack how tools like heat maps, SHAP scores, and rule extraction pull back the curtain on complex models and why these explanations are only clues about the model’s logic, not biological truth. The goal is simple: build systems clinicians can trust without surrendering judgment.
We then move through the latest research and deployments. Speech-based models for early dementia show strong accuracy, but clinical readiness remains low because workflows are ignored. On the ground, startups like Vuha Med Data are scaling cervical cancer screening in low-resource environments, while AI tools are already improving safety, communication, and clinician wellbeing through infection prediction, patient-facing chat systems, and ambient AI scribes that cut burnout within weeks.
We close with the policy and professional shifts that this rapid growth demands. New AMA requirements and India’s DPDP Act are reshaping training and data governance. And specialties like anesthesiology face an uncomfortable but necessary question: if machines handle the mechanics, how should clinicians redefine their role? The episode ends on the core challenge that will shape the decade ahead: using AI’s speed without losing the human connection that defines care.
Note: This audio podcast is an extension of The Med AI Capsule newsletter by Dr Avneesh Khare. You can explore more at www.avneeshkhare.com.
Disclaimer: This audio podcast was generated using Google NotebookLM and is based on content from the related newsletter issue. While all the information is curated from reliable sources, it is only intended for educational and informational purposes. AI LLMs can make mistakes. Please check all the important information. For any issues or inaccuracies, please reach out at [email protected].


This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit avneeshkhare.substack.com