HARC Internship Short Course

Adaptive Learning and AI-Assisted Archival Processing

A hands-on introduction for history students and archival interns.

This three-week intensive introduces students to adaptive learning models, AI-assisted archival workflows, and beginner-friendly computational methods used at the Heritage and Research Center. The course helps students understand how archives can teach systems to recognize relationships between people, places, subjects, organizations, and collections.

3 Weeks No Coding Experience Required Archives + AI Hands-On Labs HARC Collections

Central idea: students are not learning Python simply to code. They are learning how computational tools can help archivists identify relationships, recognize patterns, reconcile descriptions, and improve access to complex cultural heritage collections.

Course Format

  • Three-week short course
  • One class meeting per week
  • 2.5–3 hour sessions
  • Lecture plus hands-on lab
  • Built around HARC collections and workflows

Who It Is For

  • History students
  • HARC interns
  • Digital humanities students
  • Students interested in archives, AI, or cultural heritage
  • Beginners with little or no programming background

Primary Tools

  • Python
  • Jupyter Notebook
  • Excel and CSV workflows
  • OpenRefine
  • OCR and transcription tools
  • Simple AI-assisted workflows

Course Description

This short course introduces students to the emerging role of adaptive learning models, artificial intelligence, and computational workflows within modern archival environments. Using real collections and workflows from HARC, students learn how archival systems can be designed to identify relationships, assist with metadata generation, recognize dates and subjects, and support large-scale archival processing.

The course emphasizes beginner-friendly scripting, archival problem-solving, ethical AI-assisted practice, and user-centered access. Students explore how simple Python-based workflows and adaptive learning concepts can help archivists manage complex collections while preserving provenance and contextual relationships.

Learning Objectives

Students will learn to explain how adaptive learning models function in archives.

Students will explore how systems can recognize repeated patterns and use those patterns to improve description, identification, and discovery.

Students will understand archival relationships.

Students will examine relationships between people, organizations, places, subjects, ministries, and collections.

Students will identify common metadata problems.

Examples include inconsistent names, duplicate entities, uncontrolled terminology, fragmented descriptions, and mismatched physical and intellectual arrangements.

Students will run simple archival scripts.

Hands-on labs introduce workflows that extract names, identify dates, group related records, normalize data, and generate basic archival outputs.

Course framing: AI does not replace archivists. AI-assisted archival systems help archivists identify patterns, relationships, and inconsistencies at scales humans alone struggle to manage.

Three-Week Course Plan

Week 1 — Archives, Relationships, and Adaptive Learning

Theme: How archives teach systems to recognize relationships.

Topics

  • What is an archival information system?
  • Why archives struggle with scale
  • Metadata, relationships, and discoverability
  • Introduction to adaptive learning models
  • The Linear Reciprocity Model
  • How AI can assist archival work

Hands-On Lab

  • Explore messy spreadsheets, finding aids, and box labels
  • Identify duplicate names and inconsistent subjects
  • Map relationships between people, places, and ministries
  • Discuss how systems might learn these relationships

Week 2 — Teaching Archives to Learn

Theme: Using simple scripts to identify names, dates, subjects, and relationships.

Topics

  • Introduction to Python for archives
  • Structured and unstructured data
  • Name extraction and date identification
  • Metadata normalization
  • Controlled vocabularies and authority control
  • Pattern recognition in archival collections

Hands-On Lab

  • Run a beginner-friendly name extraction workflow
  • Clean and normalize spreadsheet values
  • Identify probable duplicates
  • Group recurring names, locations, and subjects

Week 3 — AI-Assisted Archival Systems in Practice

Theme: Building and evaluating AI-assisted archival workflows.

Topics

  • AI-assisted processing at HARC
  • Ingest and reconciliation workflows
  • EAD reconstruction and ArchivesSpace import structures
  • Box-level reconciliation
  • Ethics, bias, hallucinations, privacy, and provenance
  • Future applications in cultural heritage institutions

Final Lab

  • Complete a mini reconciliation project
  • Compare spreadsheet, box, and finding aid data
  • Normalize terms and identify relationships
  • Generate a basic archival output
  • Reflect on what requires human archival judgment

Instructional Approach

Observation

Students begin by examining real archival data, labels, spreadsheets, and descriptive inconsistencies.

Analysis

Students identify patterns, relationships, duplicate entities, missing context, and descriptive drift.

Intervention

Students use simple scripts and workflows to normalize data, extract entities, and generate outputs.

Example HARC Workflow Demonstrations

Workflow What Students Learn Archival Value
Name extraction from transcripts How systems identify people and repeated entities Improves indexing, discovery, and authority control
Spreadsheet normalization How messy legacy data becomes structured metadata Supports ingest into ArchivesSpace and Omeka S
Relationship mapping How people, ministries, places, and subjects connect Reveals hidden collection relationships
Box and EAD reconciliation How physical and intellectual arrangements are compared Preserves original order while improving access
Subject suggestion workflows How systems propose descriptive terms Supports consistent description and faceted discovery

Assessment and Reflection

Participation

Students participate in weekly labs, discussions, and relationship-mapping exercises.

Reflection Responses

Students reflect on archival systems, AI ethics, metadata problems, and the role of human review.

Mini Reconciliation Project

Students complete a small applied workflow using archival data, metadata inconsistencies, and basic normalization methods.

Potential Outcomes for HARC

This short course can support a recurring internship seminar, a digital humanities collaboration, a Notre Dame or Saint Mary’s pilot course, a computational archival science workshop, a NAPWR training framework, or a recruitment pathway into archival informatics.

Course Philosophy

Archives are not static storage environments. They are interconnected informational systems, relationship environments, evolving descriptive ecosystems, and human-centered knowledge networks.

Adaptive learning models and AI-assisted workflows help archivists manage complexity while preserving provenance, context, and discoverability. The goal of the course is not to automate archivists, but to help students understand how archivists can build systems that learn relationships while preserving historical meaning.