Skip to main content Skip to navigation

ARCHED: A Human-Centered Framework for Transparent, Responsible, and Collaborative AI-Assisted Instructional Design

Project Overview

The document presents ARCHED, a framework designed to effectively integrate Large Language Models (LLMs) into instructional design while ensuring human agency and adherence to pedagogical standards. It critiques existing AI tools for focusing excessively on automation at the expense of foundational educational principles and advocates for a structured, multi-stage workflow that fosters collaboration between humans and AI. The framework incorporates specialized AI components for generating and critically assessing learning objectives against pedagogical criteria, thereby tackling issues related to transparency, accountability, and diversity in assessment design. Empirical evaluations indicate that ARCHED successfully produces educational materials that are both pedagogically sound and well-aligned with established educational frameworks, highlighting its potential as a valuable tool in enhancing the quality and effectiveness of AI applications in education.

Key Applications

ARCHED (AI for Responsible, Collaborative, Human-centered Education Instructional Design)

Context: The framework is designed for educators in various educational settings to improve instructional design.

Implementation: It employs a multi-stage workflow that involves human educators in specifying parameters and evaluating AI-generated content.

Outcomes: The framework maintains pedagogical rigor while enhancing efficiency; it allows educators to produce high-quality learning objectives comparable to those created by experts.

Challenges: Existing AI tools often lack transparency and pedagogical foundations, risking the marginalization of human expertise.

Implementation Barriers

Transparency

Current AI systems operate as black boxes, making it difficult for educators to understand and validate the pedagogical reasoning behind generated content.

Proposed Solutions: ARCHED addresses this by separating objective generation and evaluation into distinct components that provide clear insights into the AI's decision-making process.

Automation vs. Human Agency

There is a risk that full automation will diminish the role of human expertise in instructional design.

Proposed Solutions: The ARCHED framework prioritizes human agency by involving educators throughout the instructional design process, ensuring they remain decision-makers.

Standardization

AI-generated assessments often lead to a narrow range of evaluation methods, limiting innovative learning experiences.

Proposed Solutions: The ARCHED framework encourages diversity in assessment design while ensuring alignment with learning objectives.

Project Team

Hongming Li

Researcher

Yizirui Fang

Researcher

Shan Zhang

Researcher

Seiyon M. Lee

Researcher

Yiming Wang

Researcher

Mark Trexler

Researcher

Anthony F. Botelho

Researcher

Contact Information

For information about the paper, please contact the authors.

Authors: Hongming Li, Yizirui Fang, Shan Zhang, Seiyon M. Lee, Yiming Wang, Mark Trexler, Anthony F. Botelho

Source Publication: View Original PaperLink opens in a new window

Project Contact: Dr. Jianhua Yang

LLM Model Version: gpt-4o-mini-2024-07-18