Symposium on Model Accountability, Sustainability and Healthcare
November 4-5 2025
Mila 6666 Rue Saint-Urbain, Montréal, QC H2S 3H1
The Symposium on Model Accountability, Sustainability and Healthcare (SMASH) is an interdisciplinary gathering focused on operationalizing AI safely and responsibly. The goal of this event is to identify challenges and propose technical, ethical and regulatory solutions to AI safety, data privacy, model interoperability and accountability. This symposium will explore these research topics through the lens of healthcare and sustainability.
The healthcare sector is rapidly embracing AI, and the insights gained are likely to inform future sustainability research. These disciplines operate within interconnected ecosystems of stakeholders, technologies, and datasets, sharing both the potential benefits and inherent risks.
We invite researchers to submit 4 page extended abstracts under the research areas defined below.
ML Safety, Privacy, Model Accountability and Alignment:
- Safety, robustness and alignment of ML systems
- ML systems and software traceability
- Applications of privacy enhancing technologies (PETs): differential privacy, federated learning, etc.
- ML performance and benchmarking methods
- Mechanisms for ML provenance tracking and watermarking
- Safety use cases for advanced AI systems
Interdisciplinary Submissions from Law and Social Sciences:
- Acceptance of ML technologies by clinicians, patients, and administrators
- ML risk assessments and disclosures
- Ethical considerations when deploying ML systems
- Regulation of ML technology
Contributions to SMASH that are of an applied nature are also welcome, such as case studies deploying ML with sensitive data.
Healthcare Applications:
- Applications of AI in healthcare (administration, precision medicine, patient care, diagnostics)
- Using patient data in ML research
- Healthcare ML metrics, monitoring and benchmarking
Methods for ML Sustainability:
- ML emissions accounting metrics
- Case studies and applications of sustainability in ML
- Sustainable computing and system design
- Environmental accounting methods for ML
Program Committee:
- Syed Sibte Raza Abidi, Dalhousie University
- Brian Anderson, CHAI
- Omar Benjelloun, Google
- Martin Cousineau, Obvia, HEC
- Audrey Durand, Université Laval
- Cooper Elsworth, Google
- Golnoosh Farnadi, McGill University, CIFAR, Mila
- Marie-Pierre Gagnon, Université Laval
- Jin L.C. Guo, McGill University, Mila
- Bettina Kemme, McGill University
- Eric Kolaczyk, McGill University, Mila
- Lyse Langlois, Obvia, Université Laval
- Benjamin Lee, University of Pennsylvania, Google
- Maroussia Lévesque, Harvard, CIGI
- Sumanth Ratna, CHAI, Yale University
- Charbel-Raphaël Segerie, CeSIA, ML4Good
- Carole-Jean Wu, FAIR, Meta
