Scientific Computing and Data / AIR·MS (AI Ready Mount Sinai) / News and Updates

News and Updates

NOVEMBER 2025 RESEARCH ALERT! 

Unauthorized Sharing and Access of Incoming Data with Others and Additional Compliance with Your Agreement for Data Access:

• This communication is addressed to those Principal Investigators (PIs) whose research involves accessing data from an external entity.
• Unless explicitly stated otherwise, data providers and repositories restrict data access to the individuals listed in the agreement for data access (Agreement). Your access to data does not automatically grant permission to share it with others in your lab.
• Any individual who needs access must be explicitly listed as a recipient in the agreement either by name or by inclusion in an authorized personnel category (e.g., recipient’s faculty, employees, fellows, students, and agents (“Recipient Personnel”)). It is a serious violation with potential financial and legal consequences if you share data with individuals who are not authorized. To add ISMMS individuals, follow the steps in accordance with the requirements of the data repository and the terms of your Agreement.
• The PI is responsible for ensuring that all individuals with access to the data are aware of the Agreement terms, especially any restrictions on further data access and sharing.
• External collaborators need to request access to the data repository through their own institution. The external collaborator within their institution needs to establish the Agreement.
• If you have a multiple research appointment, the data access is for you at ISMMS only. Do not send the data to your other institution or store it on their servers.
• Likewise, do not send data to a personal email account or device.
• In addition to restrictions on data sharing and access, you must ensure compliance with all other terms of the Agreement, especially those related to end of usage data destruction or return. Failure to comply may constitute a serious violation.

Events: Digital Health Partnership Workshop Potsdam 2025

“Our mission is not modest — nor should it be: We want to transform healthcare through digital innovations that directly improve people’s lives. Not someday. Not theoretically. But practically, measurably, and responsibly.” Those were the opening words of Prof. Lothar H. Wieler, co-director of our partnership, for this year’s workshop.

Setting the stage with this opening for the remainder of our exciting workshop days from November 10 through November 13, 2025, at the Hasso Plattner Institute in Potsdam, we invited research fields and departments beyond the Digital Health Cluster at HPI to present their work to our partners from the Icahn School of Medicine at Mount Sinai and Data4Life.

Our program started off with presentations on AI, cybersecurity and systems design from Prof. Falk Übernickel, Prof. Ralf Herbrich, Prof. Christian Dörr, Prof. Vaibhav Bajpai and Prof. Mathias Weske, followed by an interactive introduction to Design Thinking from the HPI d-school, a showcase of how the HPI Engine creates successful startups such as voize, Mama Health and Praxipal, as well as a tour through our HPI Maker Universe. Additionally, a big spotlight was put on our Digital Health Graduate School students to present their on-going research projects during this workshop to all three partners.

We also presented fresh developments on the Digital Discovery Program (DDP), AI-Ready Mount Sinai (AIR·MS) and SensorHub, how they will offer more opportunities going forward for students and researchers alike to enable more cross-institute collaborations using the newest research methods.

Last but not least, we heard important keynotes from our partners Dr. Nicholas Gavin, Dr. Alexander Charney and Dr. Lea Davis about the history of the Mount Sinai Health System and the Mount Sinai Millions project as a foundation for a learning health system, while Ben Illigens gave us insights into how Data4Life accelerates research and clinical translation in his keynote.

Thank you to all of our 113 participants for making this an unforgettable event! We are excited to see what we will deliver in our next Digital Health Partnership Workshop – this time in New York.

How an Open-Source Chatbot Is Accelerating Research at Mount Sinai

In collaboration with the Mount Sinai Health System and the Hasso Plattner Institute, we are building a powerful, secure, and scalable LLM infrastructure for clinical research.

Health data holds enormous potential – but only when it’s accessible, understandable, and actionable. That’s where our latest AI project with Mount Sinai comes in: an open-source, locally hosted chatbot designed to support researchers with data exploration, literature search, and documentation tasks – faster, easier, and always secure.

The Mount Sinai Health System is one of the leading medical research institutions in the world. Together with the Hasso Plattner Institute, we’re shaping AIR·MS, a secure digital research environment for structured access to health data. Now, we’re taking the next step – by integrating a Large Language Model (LLM) chatbot directly into the platform.

A chatbot built for science – not for show.

In partnership with Mount Sinai’s Scientific Computing and Data (SCD) team, we are developing an open-source LLM chatbot tailored for researchers. Unlike commercial services like ChatGPT, this chatbot runs entirely within Mount Sinai’s infrastructure – ensuring full data protection and compliance with hospital IT standards.

The chatbot supports use cases such as:

• answering clinical or technical questions,
• summarizing large volumes of literature or documentation, and
• assisting with structured health data analysis.

Secure by design – scalable by architecture.

Our chatbot, built using open-source components such as Ollama and Open WebUI, has been integrated onto the AIR·MS Application Tier. Key features include:

• Support for open weight models such as LLaMa, Mistral, and gpt-oss.
• Familiar and intuitive user interface from Open WebUI, flexible model serving from Ollama.
• MSSM Entra ID authentication and role-based access control.

Recent milestones:

• 100 initial users onboarded from >10 teams across Mount Sinai
• Security hardening through a penetration testing exercise completed with no critical vulnerabilities
• Scalable roadmap agreed upon with Mount Sinai and HPI

New Publication and Studies

“AIR·MS platform paper” – accepted and published by JAMIA Open!

We’re excited to share that our reference architecture paper on AIR·MS has been accepted for publication by JAMIA Open, the journal of the American Medical Informatics Association.

AIR·MS is a secure, scalable research platform that gives access to structured and unstructured clinical data from over 12 million patients of the Mount Sinai Health System – including EHRs, radiology, pathology, and ECG. Developed through a partnership between Mount Sinai, the Hasso Plattner Institute, and Data4Life, the platform supports international, cross-institutional research on real-world data.

The paper also highlights three real-world applications:

– AI-powered phenotyping for Crohn’s disease
– New insights into dysmenorrhea and heart disease in women
– Federated learning for cardiovascular risk prediction

Click here to learn more on our platform, use cases, and the publication