About the Curriculum, Philosophy, & Teacher

The sudden rise of AI in education has left many educators uncertain about how these tools fit into classrooms built on authorship, assessment, and accountability. This course represents a deliberate alternative to the dominant, chatbot-driven model of classroom AI.

Here, students complete their work first - designing surveys, conducting interviews, analyzing data, and writing responses - before AI is introduced at the end of the process. In most cases, the AI is run once, as an analytic instrument rather than a conversational partner.

Students remain the architects of their thinking. The AI acts as a Scientist - examining evidence, identifying patterns, and returning bounded results. The goal is not to avoid AI, but to integrate it in a way that preserves clarity, ownership, and trust in student work.

Minimal-Pass AI treats artificial intelligence as a discrete step in a workflow, not a continuous presence. It is something students run, not something they converse with.

The standard pattern in this course is simple:

1. Students produce a complete artifact (survey, dataset, transcript, or written analysis).
2. AI processes that artifact - typically once - for analysis, feedback, or synthesis.
3. The output is reviewed, questioned, and interpreted by the student.

This reflects a Scientist AI model. Unlike Agentic AI, which acts for the user and decides next steps, Scientist AI acts on student-produced material. The human defines the question, the scope, and the standards; the AI supplies computational leverage.

This approach is not merely a rejection of agents, but of unstructured AI use. While chat-based AI is simply an interface, in classroom settings it can quickly slide into agentic behavior when used without discipline, with the model implicitly guiding thinking, sequencing tasks, and shaping outcomes.

Minimal-Pass AI intentionally blocks this slide. AI is not used for idea generation, early drafts, live interviews, focus groups, or first-pass writing. Students must think, plan, and create before computation enters the picture.

This approach stands in contrast to chatbot-based classroom AI, which emphasizes continuous interaction, iterative prompting, and convenience. Minimal-Pass AI emphasizes planning, accountability, and transfer of skill beyond the classroom.

The core of this course is Social Data Science: the study of human systems using quantitative, qualitative, and mixed-methods research.

Students routinely produce:

• Quantitative surveys and instruments
• Qualitative and open-ended questions
• Interviews and focus groups protocols
• Transcripts, coding schemes, and thematic analysis
• Univariate and multivariate data analysis
• Visualizations and written synthesis

AI is applied after data collection to assist with analysis and synthesis, not to replace inquiry. AI is meant to strengthen and accelerates analysis. AI is not meant to replace thinking.

This curriculum is intentionally designed to align with multiple credentialing frameworks without altering its core structure. Flexibility comes from design strength, not dilution.

Probability & Statistics (Mathematics Credit)
Students engage deeply with data collection, variability, sampling, and interpretation. Statistical reasoning is applied to real human datasets rather than abstract exercises, aligning naturally with Statistics & Probability standards.

General Elective Credit
The course integrates research design, technical literacy, and communication. It is accessible to students from multiple academic pathways and emphasizes transferable analytical skills.

Introductory Sociology - CLEP Preparation
Course content aligns with foundational sociological concepts such as social structures, institutions, culture, and stratification. Students practice authentic sociological methods while preparing for the CLEP Introductory Sociology exam, creating a pathway to early college credit.

The CLEP (College-Level Examination Program) is a nationally recognized exam program that allows students to earn college credit by demonstrating mastery of introductory college-level material. Successful scores can translate into transferable college credit at many universities.

📄 Read the Brief Policy Report on CLEP & Louisiana SPS Alignment

The course operates on a local-first infrastructure designed around data sovereignty, focus, and instructional control. AI tools run offline within a controlled classroom network, ensuring that student work remains local, inspectable, and owned by the school community rather than external platforms.

A local-first approach treats AI as laboratory equipment rather than a remote service. Models are downloaded, run on school-owned hardware, and used for clearly defined academic tasks. This eliminates dependency on commercial subscriptions, reduces distraction, and prevents student work from being logged, trained on, or monetized by third-party providers.

• Student stations: High-performance mini PCs capable of running smaller AI models locally for structured tasks
• Teacher core: GPU-equipped systems used for heavier batch analysis, evaluation, and synthesis
• Cost: Approximately $15,000 (one-time) for a fully equipped 20-student lab

Local AI models are free, open, and downloadable, but they are intentionally smaller and less general-purpose than cloud-based systems such as ChatGPT, Claude, or Gemini. They do not attempt to replace human thinking or simulate broad conversational intelligence.

Instead, local models are used in a targeted, task-specific way. When paired with strong structure, clear prompts, and defined academic standards, smaller models can be highly effective at analysis, transformation, and feedback on student-produced work. Their limitations become a feature rather than a liability, reinforcing disciplined use and preventing overreliance.

While the full lab uses dedicated hardware to support this model at scale, the curriculum itself is adaptable to standard Chromebooks or shared devices. The instructional philosophy remains the same regardless of hardware: students think and create first, computation enters deliberately, and AI remains a tool rather than an authority.

Seán Muggivan is a New Orleans-based educator and the creator of the Muggs of Data Science lab. With over a decade of classroom experience, preceded by 5+ years of practice in social work, his work is shaped by sustained, firsthand engagement with students and the real constraints of school environments.

The Muggs of Data Science lab integrates statistics, sociology, and applied artificial intelligence into a single instructional framework. Rather than teaching tools in isolation, students investigate real human systems through surveys, interviews, datasets, and written analysis, producing work that is visible, assessable, and grounded in authentic academic practice.

Muggivan’s approach reflects what has proven durable in the classroom over time. His background in mathematics education, special education, and social work informs a systems-level view of teaching and technology, where structure, clarity, and accountability matter as much as innovation.

He champions Scientist AI: using models to act on student-produced work rather than to act for students. In this framework, teaching is not content delivery but a living laboratory, where tools, boundaries, and learning outcomes are continuously tested against classroom reality.

The Muggs of Data Science lab and its local-first AI approach have been featured in the local news media.

• Fox 8 Live: AI comes to a New Orleans charter school
• Television interview segment
• Classroom feature and discussion