In this episode, Andy Slavitt and the hosts discuss the rapid adoption of AI in healthcare, emphasizing that leaders must stay actively engaged as AI tools like LLMs are increasingly integrated into daily workflows by both clinicians and patients. They highlight the opportunities and challenges in regulation, organizational culture, and patient empowerment, concluding that AI’s transformative potential in healthcare is accelerating and likely to bring significant benefits despite some risks.
In this episode of the Stanford Healthcare AI podcast, hosts Justin and Matt are joined by Andy Slavitt, co-founder and general partner at Town Hall Ventures and former senior advisor to the Biden administration for pandemic response. The discussion opens with an overview of the current state of AI adoption, particularly large language models (LLMs) like Gemini, ChatGPT, and Claude. The hosts note rapid progress in AI capabilities, especially in image generation and text, and highlight how users are developing preferences for specific models. Despite the fast pace, there is still a sense that the field is wide open, with no single model dominating all use cases.
Andy Slavitt shares insights from his experience at Town Hall Ventures, explaining how his team has adopted Gemini for work due to its integration with Google Workspace, even though many team members personally prefer ChatGPT. He describes the evolution of their approach to AI: from struggling to identify promising healthcare AI investments, to realizing their own portfolio companies—and even their own team—needed to more fully embrace AI. To stay at the forefront, they are hiring an AI research fellow to guide adoption and keep the team updated on the latest advancements.
A key theme of the conversation is that AI adoption in healthcare cannot be delegated or treated as a one-time decision. The technology is evolving too quickly for leaders to remain hands-off; instead, they must stay engaged and foster a culture where AI is integrated into daily workflows. The hosts and Andy discuss how, unlike previous healthcare technologies that required top-down enforcement, AI is being pulled into organizations by users who find it genuinely helpful. This bottom-up adoption is creating a disconnect, as many clinicians use AI tools daily but feel unsupported or even restricted by their organizations.
The panel addresses the regulatory and policy landscape, noting both the challenges and opportunities. Andy highlights recent policy signals, such as the new CMS Access Model, which incentivizes technology companies to improve patient outcomes, not just pay physicians for care. He also points out that the FDA is streamlining approval processes for AI tools, signaling strong government support for rapid AI integration in healthcare. However, Andy cautions against overregulation, arguing that the pace of AI development makes traditional regulatory approaches difficult and potentially counterproductive.
Finally, the conversation turns to the broader implications for patients and the healthcare system. The hosts emphasize that AI is empowering not just clinicians but also patients, especially those with limited access to care. Tools like LLMs can help patients better understand their health and prepare for medical visits, potentially reducing disparities. While there are concerns about monetization models—such as advertising in AI tools—and the risk of negative media coverage focusing on rare failures, the overall sentiment is optimistic. The panel agrees that AI is at a transformative moment in healthcare, with adoption accelerating at both the consumer and enterprise levels, and that the benefits are likely to outweigh the risks as the technology matures.