GenAI‑Enhanced Learning for Biosensing and Microfluidic Analysis

Generative artificial intelligence (GenAI) is rapidly reshaping how students learn, problem‑solve, and communicate scientific concepts. In biomedical engineering education, there is a need for learning activities that connect foundational biomolecular theory with real‑world applications while developing responsible AI literacy. This project enhances BME2106 – Introduction to Cellular and Biomolecular Engineering by introducing a GenAI‑integrated, problem‑solving mini‑project centered on biosensing and microfluidic analysis. The redesigned activity offers students a safe, structured, and realistic opportunity to practise applying core BME2106 concepts—such as tissue and cell heterogeneity—to simplified experimental systems using pre‑fabricated microfluidic chips and fluorescence‑based assays from lab sessions. This aligns with the course’s emphasis on understanding cellular microenvironments and biochemical characterisation techniques. Students work through authentic, real‑world problem scenarios modelled after applications found in diagnostics. Examples include predicting how changes in fluorescence reflect binding interactions relevant to biosensor technologies. These scenarios mirror practical applications encountered in contemporary biosensor systems, allowing students to experience how foundational biomolecular engineering concepts translate into measurable signals used in healthcare and biotechnology. To support learning and AI literacy, the project incorporates four structured GenAI checkpoints. First, during concept exploration, students use the CityUHK GPT Chatbot to clarify biosensor mechanisms, explore examples of real‑world biosensing systems, and visualise fluorescence signalling. Second, in the design‑reasoning phase, GenAI helps students compare and justify choices among instructor‑provided microfluidic layouts based on expected biological behaviour, supporting structured decision‑making while maintaining feasibility for undergraduates. Third, after completing short, safe hands‑on activities involving solution loading and basic imaging, students use GenAI for first‑pass interpretation of fluorescence patterns and then refine, correct, or reject AI‑suggested explanations, reinforcing analytical judgment and critical evaluation. A final GenAI checkpoint supports scientific communication. Students use GenAI to draft poster layouts, generate illustrative schematics such as diffusion diagrams or microchannel graphics, and design figure panels. Under guidance, they modify and validate all AI‑generated visuals to ensure conceptual accuracy and proper attribution. This stage strengthens communication skills while modelling responsible AI‑supported visualisation practices. Students work in small groups during practical components but complete an individual reflection documenting their problem‑solving process, biological reasoning, and AI‑use evaluation. Assessment includes a poster presentation and a rubric‑based evaluation of AI critique and data interpretation. Poster templates will be disseminated through the Digital CityU Collections TDG repository for wider adoption across STEM courses.This project strengthens students’ problem-solving confidence and readiness for real-world biomedical applications. By integrating GenAI into early-stage reasoning and late-stage communication, students gain practical experience in how AI-augmented workflows increasingly support tasks in diagnostics, biomedical R&D, and data-driven healthcare innovation. The structured checkpoints ensure that GenAI serves as a scaffold rather than a shortcut, enabling students to practise essential skills such as identifying experimental limitations and refining interpretations based on both empirical observations and AI-generated insights. Overall, this module provides a scalable model for embedding GenAI into laboratory-linked teaching, promoting reproducibility, reflective practice, and effective scientific communication.