Below is the AI-generated summary of the event, with minor edits made by us:


Quick recap


The intersection of AI and healthcare was discussed, with a focus on the shift from academic to industrial research and the potential benefits of both predictive and generative AI in healthcare practice. The career trajectory of Dr. Suchi Saria, a healthcare AI expert, was highlighted in a conversation with Drs. Tinglong Dai and Risa Wolf, emphasizing the importance of interdisciplinary collaboration, patient-centered care, and continuous learning. The conversation also touched on the need for a data-driven approach to implementing AI in healthcare systems, the challenges of regulation and compliance, and the importance of responsible data sharing and change management.




AI and Healthcare: Research Shifts and Predictions

Drs. Tinglong Dai, Suchi Saria, and Risa Wolf, three experts in healthcare AI at Johns Hopkins, discussed the intersection of AI and healthcare, highlighting the shift of AI research from academia to industry and the benefits of this transition. Dr. Saria, a leader in the Johns Hopkins Malone Center for Engineering and Healthcare, was introduced and her AI Healthcare Lab was highlighted for its contributions to predictive and generative AI. The discussion also touched on the need for a comprehensive approach to building AI healthcare solutions and the potential of generative AI in healthcare practices. Suchi's career path and the dynamic relationship between academia and industry were also discussed.


Suchi's AI Journey and Healthcare Improvements

In a conversation with Risa and Tinglong, co-chairs of the HBHI Workgroup on AI and Healthcare, Suchi shared her passion for artificial intelligence since childhood and her role in developing AI systems at major tech companies including Facebook, Meta, Apple, and Stanford. She discussed her vision for data infrastructure and the potential of AI to solve complex, long-term problems through interdisciplinary collaboration. She also highlighted the challenges of translating AI into practical applications and emphasized the importance of AI in improving healthcare by predicting complications and reducing documentation burden.


Scalable Solutions in Healthcare Impact

Continuing her conversation with Risa and Tinglong, Suchi discussed the challenges of making an impact in medicine and the need to learn from the best in order to develop scalable solutions. She shared her career trajectory, which has focused on advancing electronic health record data techniques and developing a clinical AI platform. She highlighted her work at the FDA and the creation of a spin-off company, Bayesian Health, which aims to identify patients early for various diseases using AI-driven care models. Suchi also discussed the importance of balancing a career in academia with running a company, and the role of patience in problem solving. She touched on her past experiences at Stanford and how the technology and education systems in Silicon Valley have influenced her career.


AI Technology and Risk Assessment Discussion

Suchi discussed her experiences as an entrepreneur and translator, highlighting her observations about the limitations of electronic medical record (EMR) data. She highlighted the potential benefits of AI technology to address issues with noisy data. She also suggested the need for a practical framework to promote transparency, especially when considering patient value.

Dr. Ray then initiated a discussion on risk assessment and regulatory frameworks for AI, broadening the conversation beyond the clinical domain. The group agreed on the need to consider risk across all AI applications, not just in healthcare.


Transformative Technologies, Feedback Loops, and IRBs

Suchi emphasized the importance of advancing transformative technologies that improve healthcare. She discussed the concept of feedback loops in healthcare and how robust learning algorithms can be developed to avoid them. She also clarified the role of an Institutional Review Board (IRB) in deploying solutions, advocating for risk-benefit analysis, and studying outcomes. Finally, she suggested that the FDA should be considered for validation if the goal is to scale the solution nationally and make marketing claims about it.


Advanced AI and Health Informatics Teaching

Suchi shared her experience teaching advanced AI and health informatics courses and emphasized the importance of regularly updating course content. She noted that this approach not only keeps the curriculum relevant, but also helps students appreciate the rapid evolution of fields like computer science and engineering. Suchi advocated the importance of continuous learning and the need to create innovators who can think differently and create with data. She also expressed concern about resistance to change among busy professionals and the need for better change management strategies.


Data-Driven AI in Healthcare Discussion

Suchi discussed the importance of a data-driven approach to implementing AI in healthcare systems. She emphasized the need to monitor system performance, continuously tune solutions, and understand organizational culture when implementing new technologies. She also expressed interest in collaborating with others working in this area. Stuart expressed concern about the challenges of continuous learning systems in the FDA regulatory environment and asked how AI solutions could address this issue.


Improving Systems, Policies, and Human-Machine Teams

Suchi and Stuart discussed the importance of bringing together best practices and expertise to improve systems and policies, with a focus on learning from data and industry examples. They also addressed concerns about commercial use of data without consent, the need for responsible data sharing, and managing changing patient inputs into the system. Nestoras raised a question about the integration and testing of predictive models in an EHR system and the timing of potential FDA regulation. Harold highlighted the changing epistemology in medicine and the need for updated education and training for medical students. Vijaya emphasized the importance of monitoring and nurturing human-machine teams to foster openness to change and adaptation.