ACM TRUST 2027

Track Description

Responsible, Explainable, and Ethical-by-Design Artificial Intelligence in Clinical Applications

Track Scope

From Trustworthy Methods to Clinical Deployment

REED-AI advances trustworthy AI methodologies and deployment frameworks for healthcare and clinical environments. It welcomes original research, position papers, and case studies across explainable and interpretable AI, uncertainty quantification, fairness and bias mitigation, multimodal and foundation models, clinical natural language processing, medical imaging AI, federated and privacy-preserving learning, human-AI collaboration, governance, and post-deployment monitoring.

Particular emphasis is placed on generalizability, reliability, clinician trust, workflow integration, and societal impact.

Motivation

Healthcare Is a High-Stakes Test of Trust

Foundation models, multimodal AI, and generative AI create valuable opportunities in medical imaging, clinical decision support, healthcare operations, and patient engagement. Their clinical deployment remains constrained by concerns about explainability, robustness, uncertainty, fairness, accountability, privacy, and governance.

REED-AI aligns with ACM TRUST 2027 by addressing the complete ecosystem required for trustworthy adoption, including human-centered design, regulatory considerations, transparency mechanisms, uncertainty-aware decision support, fairness evaluation, privacy-preserving learning, and continuous monitoring.

Learning Objectives

  1. Trustworthiness challenges involving transparency, uncertainty, fairness, accountability, privacy, and patient safety.
  2. Responsible and explainable methods across medical imaging, clinical language technologies, multimodal learning, and foundation models.
  3. Ethical-by-design clinical AI systems developed through interdisciplinary and human-centered approaches.
  4. Validation, deployment, monitoring, governance, privacy, safety, and regulatory strategies for translation.
  5. Perspectives from computer science, healthcare, digital health, policy, ethics, and design.
  6. Research gaps and future opportunities for robust, equitable, transparent, and clinically impactful AI.

Program at a Glance

01

Scientific Session

Approximately 8-10 presented papers spanning mature research, emerging work, positions, posters, and demonstrations.

02

REED-AI Hackathon: Explainable AI in Medical Imaging

Teams will evaluate predictive performance, explanation quality, uncertainty, failure modes, and clinical communication.

03

Invited Talks and Panel

Two invited talks, a panel of 3-4 experts, open discussion, and cross-institutional networking.

Sustainability

A Long-Term REED-AI Community

The organizers envision a recurring forum supporting mentorship, collaborative publications, shared benchmarks, dataset challenges, and cross-institutional research. Lessons from the scientific session, hackathon, and community discussion will inform future recommendations and research directions.