# bok-ai-lab-20250404

## **AI Lab Report – Week 2: Polling as Pedagogical and Computational Practice**

In our second session, we delved into the practice of classroom polling, examining it from both pedagogical and computational lenses. We explored potential AI-driven interventions to enhance polling processes within General Education courses, driven by the central question: *How can AI enhance polling activities in a General Education course?*

To initiate discussions effectively, we reviewed pertinent AI news, including Sean Kelly's recent AI panel. We utilized foundational definitions and a glossary to support diverse participant backgrounds, fostering a common understanding of key concepts and facilitating inclusive engagement.

### Key Experiments and Discussions:

#### 1. Human Web Polling Activity
We conducted an interactive exercise designed to model traditional polling mechanics:

- Participants requested and received paper ballots from Madeleine, the facilitator.
- Ballots were filled out individually and returned to Madeleine, who tabulated votes manually.
- Madeleine created and shared a summary card, visualizing collective responses.

In an enhanced iteration, Madeleine collaborated with an assistant to manage more complex tabulation, demonstrating scalability and complexity in polling scenarios.

This exercise illuminated the step-by-step mechanics of polling, identifying clear opportunities for intervention into this workflow with AI.

#### 2. AI-Enhanced Polling Prototyping and Testing
Participants explored multiple experimental approaches leveraging AI:

- **Paper Prototyping:** Rapid, iterative development of user interfaces and interactions, allowing immediate feedback on conceptual designs and user experiences.
- **Vision-Driven Data Capture:** Python notebooks demonstrating the use of camera vision technology to automate the capture and digitization of handwritten polling inputs.
- **CSV Data Management:** Using Python notebooks, we explored parsing CSV files to streamline data processing and facilitate efficient analysis of poll responses.
- **Front-End Interface Prototyping:** Creation of responsive polling interfaces, integrating AI-generated suggestions and results visualization, prototyped in Next.js and supported by Claude artifact front-end tools.

### Insights and Methodological Reflections:

- **Reconceiving Polling as Computational Querying:** By reframing polling activities as iterative computational tasks (such as repeated HTTP requests), we opened avenues for innovative AI integration, emphasizing how pedagogical processes can benefit significantly from computational analogies and automation.

- **Mapping AI Integration Points:** We systematically analyzed existing polling workflows, identifying precise points at which AI interventions could meaningfully augment and optimize polling processes—from initial query to final analysis and feedback delivery.

- **Platform and Technical Foundations:** We emphasized foundational web concepts, covering client-server dynamics, React and Next.js frameworks, headless CMS options, HTML/CSS fundamentals, and basic HTTP requests. These elements formed the technical backbone necessary for participants to understand, conceptualize, and implement sophisticated AI-driven polling solutions.