Artificial intelligence models are increasingly developed for use in clinical contexts, including diagnosis, risk prediction, treatment planning, and patient interaction. Alongside their potential benefits, such systems raise substantial ethical challenges, particularly with respect to fairness, bias, accountability, transparency, privacy, and the impact of AI-supported decisions on patients and clinical workflows.
In clinical settings, ethical concerns are tightly intertwined with the robustness and resilience of AI systems to data shifts, noise, adversarial conditions, and deployment-time uncertainty. Failures in algorithmic resilience can disproportionately affect vulnerable patient groups and undermine trust, accountability, and patient safety. At the same time, the increasing use of synthetic data to address data scarcity, privacy constraints, and representation gaps introduces new ethical questions regarding bias propagation, fidelity, and downstream clinical validity.
We welcome submissions from across AI disciplines, including machine learning, natural language processing, computer vision, multimodal systems, and foundation models, as applied to clinical settings.
Topics of interest include (but are not limited to):
- Fairness and bias in clinical AI systems
- Ethical challenges in AI-based clinical decision support
- Robustness and resilience of clinical AI systems, e.g., robustness to data distribution shifts, missing or corrupted data, adversarial inputs, and real-world deployment conditions
- Representation gaps, data imbalance, and under-served patient populations
- Bias and uncertainty in multimodal and foundation models for healthcare
- Accountability, explainability, and contestability in clinical AI
- Ethical evaluation beyond accuracy: metrics, benchmarks, and validation practices
- Human–AI interaction and shared decision-making in medicine
- Privacy, consent, and secondary use of clinical data
- Ethical implications of synthetic data generation in healthcare, e.g., fairness, representational fidelity, privacy guarantees, bias amplification, and evaluation of models trained on synthetic data
- Stress testing, auditing, and failure analysis of clinical AI models
- Regulatory, legal, and governance challenges for clinical AI
- Case studies of real-world clinical AI deployments and failures
- Methodological frameworks for ethical-by-design clinical AI
Submission Types
We invite two three types of submissions:
- Long papers (8 pages): completed and original research contributions
- Short papers (4 pages): work in progress, position papers, or preliminary results
- Extended abstracts (1 p, ref not included): projects, work in progress, etc
Depending on your choice, submissions can be either:
- Archival, meaning they present original, unpublished work that will be included in the workshop proceedings, or
- Non-archival, meaning they may report on previously published work, work under review elsewhere, or ongoing research, and will not be included in the proceedings
All submissions will be peer-reviewed. Accepted papers will be presented at the workshop as either oral presentations or posters.
Authors should use the IJCAI author kit: https://www.ijcai.org/authors_kit
Submission site: https://cmt3.research.microsoft.com/ETHICAIA2026
Outstanding contributions may be invited to a journal special issue (to be confirmed).
The Microsoft CMT service was used for managing the peer-reviewing process for this conference. This service was provided for free by Microsoft and they bore all expenses, including costs for Azure cloud services as well as for software development and support.