Mayo Clinic AI Summit

Introducing the 2026 AI Summit!

We are pleased to invite you to participate in the AI Summit conference titled “AI Summit: A New Engine for Real-World Evidence Generation: Multi-Agentic AI and Simulation” to be held in Rochester, Minnesota on June 4-5, 2026. The conference is organized by the Mayo Clinic Department of Artificial Intelligence and Informatics in partnership with the Mayo Clinic AI community.

The rapid advancement of large language models has transformed how medical evidence can be captured, interpreted, and generated. Building upon this foundation, multi-agentic AI enables intelligent coordination and reasoning across complex, data-rich tasks—amplifying the capacity of AI systems to manage scale and complexity. Through simulation frameworks, these models can move beyond analyzing what has been observed to reasoning about and generating evidence for scenarios that have not yet occurred. Together, multi-agentic AI and simulation establish a new paradigm for real-world evidence generation—linking data, reasoning, and virtual experimentation to create insights that are both generalizable and clinically meaningful.

You are invited to submit proposals to be considered for presentation through the conference registration and submission website. 

Registration will open January 2, 2026.

Abstract Submission:

We invite the submission of abstracts to be considered for the following presentation types and topic areas:

The abstract submission window will open January 2, 2026.

Presentation Types:


 Lightning Talk (10 minutes)

 Poster Presentation

 Workshop/Tutorial (up to 3 hour blocks available)

Publication Opportunity:

Top abstracts will be selected for publication in Mayo Clinic Proceedings - Digital Health

Topics:

1.    Multi-Agentic AI as Clinical Evidence Engines

  • Autonomous Orchestration for CDS — Learning-based agent coordination that infers context, selects tools/policies, and resolves conflicts with human-in-the-loop safeguards.
  • Decentralized Processing, Centralized Reasoning — Local/federated multimodal processing unified by a centralized causal/graph/Bayesian layer for auditable decisions.

 
2.    Simulation Frameworks for Medical Data

  • Trial Emulation & Scenario Simulation — RWD-driven protocol emulation with counterfactual and policy analyses to stress-test care pathways and interventions.
  • Digital Twins — Patient- and cohort-level models rigorously calibrated and externally validated to close the sim-to-real gap.
  • Digital Animal/Cell for Mechanism & Toxicity — Biological data–driven virtual organisms/cells for mechanism exploration and pharmaco-tox prediction, packaged with regulatory-ready reporting.

 
3.    LLM-Based Foundational Models for RWE

  • Multimodal LLM Training & Fine-Tuning — Train and adapt multimodal LLMs on literature, EHR notes, imaging, and physiologic signals to deliver executable analytics and power downstream predictive tasks.
  • Data & Patient Searching — Semantic retrieval and cohort/patient matching with transparent selection criteria for research and recruitment.
  • Responsible Deployment (Equity & Community) — Transportability and fairness checks using socioeconomic/geospatial factors, continuous drift/safety/cost monitoring, and clinician/patient co-design.