BP801T · Core · Semester 8
Ethical Considerations & Translational Applications of AI in Pharmacy
Study AI lifecycle, validation, auditing, regulatory and governance concepts, pharmacy automation, public-health data concepts, and responsible translation through non-operational educational scenarios.
Course objectives
- Introduce the lifecycle of AI systems used in pharmaceutical and healthcare applications, including data management, development, validation, and deployment concepts.
- Provide understanding of model validation, auditing procedures, and documentation practices for reliable and reproducible AI systems.
- Familiarize students with regulatory, governance, and ethical frameworks guiding AI implementation in pharmaceutical and healthcare environments.
- Explore AI applications in pharmacy automation, supply-chain management, and public-health data analytics as bounded concepts.
- Develop practical understanding of AI implementation in pharmaceutical and healthcare domains through case studies and guided project work.
Unit 1: AI Lifecycle, Validation & Model Auditing
AI lifecycle teaching covers fictional data collection, preprocessing, modeling, validation, deployment, monitoring, data quality, train/test/validation splits, conceptual cross-validation, model drift, data leakage, documentation, reproducibility, and auditing. This activity does not manage healthcare data, validate a real model, or authorize deployment.
Choose a fictional lifecycle stage, identify a leakage or drift risk, and explain why documentation and auditing cannot be replaced by a local score.
Unit 2: Regulatory Framework & Explainable AI
Regulatory-submission, explainable-AI, transparency, interpretability, accountability, risk-based classification, and EU AI Act/FDA/CDSCO framework vocabulary can be studied through a fictional governance lens. This activity is not regulatory guidance, a submission, a classification, legal advice, or a compliance determination.
Inspect a fictional evidence gap, distinguish transparency from interpretability, and state why a model output cannot establish compliance or accountability.
Unit 3: AI in Pharmacy Automation & Supply Chain
Pharmacy-automation and supply-chain teaching covers automated dispensing concepts, inventory prediction, regression-based demand forecasting, medication-adherence monitoring vocabulary, supply-chain risk, logistics, privacy, legal considerations, and limitations. The activity cannot operate dispensing, forecast inventory, monitor adherence, or make logistics decisions.
Use a fictional automation workflow, identify a forecast or monitoring limitation, and name a legal, privacy, or human-oversight constraint.
Unit 4: AI in Public Health & Real-World Data Analytics
Public-health and real-world-data teaching covers fictional EHR, claims, and surveillance structures; outbreak, population-level risk, epidemiology, vaccination, prevalence, and antibiotic-resistance trend concepts. No real records are accessed, no individual risk is assessed, and no public-health forecast or intervention recommendation is produced.
Map a fictional data source and population pattern, then explain why representativeness, privacy, and uncertainty prevent individual or operational conclusions.
Unit 5: Guided Project — Translational AI in Pharmacy
A guided translational project frames fictional formulation, PK, ADR, quality-control, automation, or public-health data concepts; a supervised model and validation plan; regulatory implications; bias and privacy risks; and a structured peer/faculty-review implementation plan. The canvas does not train, validate, or deploy a real model or process real-world pharmacy data.
Build a fictional implementation plan, name a validation, regulatory, bias, or privacy question, and write a no-deployment statement for peer or faculty review.