Human-AI Interaction in Medicine: Risk Prediction and Regulatory Perspectives

Context
Artificial intelligence (AI)-driven risk prediction models are increasingly used in medicine, from cancer prognosis to cardiovascular risk scoring. However, their integration depends on responsible design, clinician-AI interaction, and regulatory compliance. This project focuses on bridging technical model development with policy, ethics, and legal frameworks that govern AI in healthcare.
Research Objectives

• Risk Prediction Models: Train and validate deep learning/large language models using medical imaging and molecular datasets.
• Human-AI Interaction Studies: Analyze how clinicians/healthcare professionals interact with AI systems and how trust is built (or broken).
• Regulatory Alignment: Evaluate how upcoming policies (e.g., EU AI Act, FDA frameworks) impact AI adoption.
• Safety & Accountability: Implement mechanisms to enable patient safety and clarify liability in AI-assisted care.

Programs & Tools Students Can Use

• Deep Learning Frameworks: Keras, PyTorch, TensorFlow for imaging and molecular data models.
• Medical Imaging Libraries: MONAI (Medical Open Network for AI), SimpleITK, 3D Slicer for healthcare datasets.
• Data Management: HDFS, PostgreSQL, FHIR-based APIs for integrating electronic health records.
• Explainability & Trust Audits: InterpretML, and Counterfactual Explanations Toolbox.
• Regulatory Simulation Tools: Compliance checklists based on EU AI Act, FDA SaMD (Software as a Medical Device) frameworks.
• Qualitative Tools: NVivo, Atlas.ti for coding clinician interviews on trust & regulatory perception.

Expected Outcomes

• Development of explainable, regulation-ready risk prediction models.
• A comparative study of global regulatory frameworks in AI-driven medicine.
• Policy briefs and academic papers supporting trustworthy clinical AI.

Student Gains
Students will gain cross-domain expertise in AI model development, regulatory analysis, and ethical AI design. They will work at the intersection of technical innovation and health law, preparing them for leadership roles in AI governance in medicine. Students will also train models on open medical datasets, simulate regulatory approval pipelines, and write structured regulatory compliance reports, bridging technical and policy insights.
Requirement
Proficiency in Python and one deep learning library (PyTorch, TensorFlow, or Keras) is mandatory. Experience with medical imaging libraries (e.g., MONAI, SimpleITK) or prior knowledge of regulatory frameworks (EU AI Act, FDA SaMD) will be a strong asset.
Related Areas of Project
[Computer Science and Engineering, Molecular Biology and Bioengineering, Medical Imaging and Vision Science, Artificial Intelligence and Deep Learning, Policy and Regulatory Studies, Ethics in Technology]
Application Process (Final Note):
You should prepare:
Latest Academic Transcript (official or system-generated PDF from your student portal)
1-page Letter of Interest explaining:

• why you are interested in the project,
• any prior experience (coursework, projects, internships),
• how you want to contribute (e.g., technical development, policy analysis, data science).

After you complete your online application, email the documents above to polat.goktas@sabanciuniv.edu with the chosen Project Name in the subject line.