An Artificial Intelligence Powered Future
The 2024 School of Medicine Leadership Retreat was a whirlwind of groundbreaking ideas, passionate discussions, and a collective leap forward in our understanding of AI's potential in research, education, and clinical care.
We explored the vast possibilities of AI-powered diagnostics, personalized learning tools, and intelligent assistants. We grappled with the ethical considerations, need for responsible development, and crucial role of human oversight in developing and maintaining trust as we move forward.
And through it all, the dedication and enthusiasm of our community shone through, leaving me confident that we are on the cusp of something truly transformative.
However, our journey doesn't end here. This retreat has been a springboard, igniting the spark of collaboration and propelling us towards a future where AI integrates into the fabric of all we do at UCSF. As we think to the near future and implementation of AI in our mission areas, I urge all of us to consider:
Embracing collaboration;
Focusing on the human element;
Prioritizing responsible development; and,
Investing in education and training.
The future of research, education, and clinical care, shaped by AI, is brimming with potential. However, it's up to us to ensure it's a future that benefits all, a future where technology empowers, uplifts, and improves the lives of countless individuals.
Sincerely,
Talmadge E. King, Jr., MD
Dean, School of Medicine
Artificial intelligence is revolutionizing
the way we approach research, education and clinical care.
Aylin Caliskan, PhD
Peter Lee, PhD
Stuart Russell, PhD
Aylin Caliskan
Assistant Professor, The Information School
Paul G. Allen School of Computer Science & Engineering (courtesy) Co-director • Tech Policy Lab
Faculty Affiliate • UW NLP RAISE, VSD Lab
University of Washington
Read her bio
What we learned:
Current LLM models of AI transfer biased information from human society to AI by reflecting the statistical regularities of language. Bias is also driven by underrepresentation of groups and broad categorization that erase minority group nuances.
Generative AI models such as text-to-image generators amplify bias.
Addressing bias in generative AI will require technical approaches, policies and awareness among researchers, policymakers and the public. We need to be able to audit the “black box” that is generative AI.
Stuart Russell
Michael H. Smith and Lotfi A. Zadeh Chair in Engineering
Professor in the Division of Computer Science, EECS
UC Berkeley
Read his bio and his book: "Human Compatible"
What we learned:
The purpose of artificial intelligence is to fulfill human objectives, and we must avoid the standard model of AI, in which machines are designed to carry out their goals regardless of the human cost. These ideas form the central theme of Dr. Russell’s book, Human Compatible: Artificial Intelligence and the Problem of Control.
An ideal approach to general AI is knowledge-based, in which the machine is capable of deliberation and reasoning on the basis of learned knowledge and human benefits.
Current neural network-based AI systems, such as large-language models, are powerful but unlikely to be the final solution to general AI, due to their intrinsic and unavoidable flaws, including their tendency to adopt internal goals that we do not understand or control. Some hybrid of well-founded AI and deep learning is likely to be the best approach.
Peter Lee
Corporate Vice President of Research and Incubations, Microsoft
Read his bio and his book "The AI Revolution in Medicine: GPT-4 and Beyond"
What we learned:
GPT-4 shows great potential for reducing clinician time spent on administrative and medical record documentation tasks in health care.
Using voice recognition platform provides a more natural, iterative engagement with GPT-4, and eliminates the need to type questions into a keyboard.
GPT-4 has an inherent need to provide the user with a convincing answer, even when it's manufactured and false. However, you can also use GPT-4 to confirm the accuracy of responses, and to identify bias in responses.
Enabling GPT-4 to use tools, for data retrieval, data analysis, and computer programming, opens broader possibilities for applications in medicine.
Background Reading
"Artificial Intelligence - from starting pilots to scalable privilege"
"Foundation models for generalist medical artificial intelligence"
"AI in Medicine: Benefits, Limits, and Risks of GPT-4 as an AI Chatbot for Medicine"
"Effective gene expression prediction from sequence by integrating long-range interactions"
"AI recognition of patient race in medical imaging: a modelling study"
A Poster Session and multiple Flash Talks
offered insight on AI applications at UCSF
Poster Session:
Click here to view posters from our talented faculty, staff and learners that showcase
how we can harness the power of AI to advance our mission areas.
Flash Talks:
We heard from 15 colleagues about how they are experimenting with and using AI.
Exploring Where the SOM Can Lead
In the rapidly evolving space of AI, much of the large investment
will be done by the private sector in refinement and development of new models and products.
Our retreat participants focused on areas where the SOM could lead in the application and evaluation of these tools across our mission areas; for example, by standing up systems and structures to support AI experimentation and implementation and ensuring ethical and equitable implementation of AI.
Ethics and Bias
Develop a strategy for UCSF to become a national leader in AI for health equity—as through research, education, patient care, and advocacy. Include equity and ethical use of AI as a primary goal/mission, with accountability and redress
Education
Update the “physician of the future” – to include new competencies related to AI.
Clinical Care Delivery
Develop a strategy for SOM-wide adoption of emerging AI technologies in clinical care. Establish an algorithmic vigilance system for initial evaluation and longitudinal impact monitoring of AI tools. Create an incubator for innovative AI ideas.
Administration
Launch a process to identify the areas that have the greatest potential of impact via AI (i.e. small, focused project), leveraging the work we are doing with the ERP and by other campuses.
Translational Clinical Research
Strengthen pipeline from discovery to clinical evaluation. Including EHR sand box/test bed; transparency of processes; implementation science framework; and support for AI integration.
Discovery
Develop educational plan for discovery science faculty and trainees in the uses and opportunities for the application of AI to their scientific questions.
The discussions and output from the Leadership Retreat will serve as the building blocks for the School of Medicine's 2025-2030 strategic plan. We will continue our conversations in these key areas and further explore how the School of Medicine can make an impact in this space to ensure that AI is accessibly, ethically and equitably implemented.