# ARCHED: Human-Centered AI-Assisted Instructional Design

Canonical page: https://yiziruifang.com/project/arched/
Website owner: Yizirui Fang

A human-centered framework for transparent, responsible, and collaborative AI-assisted instructional design.









## Sources and artifacts



- [Paper](https://yiziruifang.com/publication/li-2024-arched/)

- [Proceedings](https://proceedings.mlr.press/v273/li25a.html)

- [PDF](https://raw.githubusercontent.com/mlresearch/v273/main/assets/li25a/li25a.pdf)

- [OpenReview](https://openreview.net/forum?id=AtpVLle6G3)





Published in the **iRAISE 2025 workshop proceedings**, PMLR 273, pages 94–104. The [publisher's record](https://proceedings.mlr.press/v273/li25a.html) and [paper PDF](https://raw.githubusercontent.com/mlresearch/v273/main/assets/li25a/li25a.pdf) provide the full method and author list.

## Research question

How can LLMs support instructional design while preserving educator control and alignment with learning objectives?

## Method

ARCHED separates a Learning Objective Generation System from an Objective Analysis Engine. Educators specify the teaching context, inspect Bloom's taxonomy classifications, and select or refine objectives before assessment development.

## Evaluation and scope

The evaluation examines learning-objective quality, taxonomic alignment, and preliminary feedback from educational experts. These results concern support for the instructional-design workflow; they do not establish classroom learning gains or broad adoption outcomes.



---

[All papers](https://yiziruifang.com/publication/) · [Research](https://yiziruifang.com/research/)