A Global Review of AI in Assessment Design

By Digital Education Council

July 7, 2025

01 AI First, Human Revision

Description

Students begin the assessment by using a generative AI tool to produce a first draft or solution. The student then takes over to evaluate, revise, and build upon the AI-generated content.

Example

Editing an AI Essay with Tracked Changes

1. Students prompt an AI to write a 500-word essay on a familiar topic, intentionally producing a version that is factually inaccurate, poorly written, or stylistically weak.

2. Using track changes in a word processor, students edit the essay to improve clarity, accuracy, and logic.

3. Students annotate each revision with brief explanations and submit a one-page addendum describing the prompts used, common issues found, and reflections on AI writing behaviour.

4. Students submit the original essay, edited version, and addendum.

PROMPT:

Write a 500-word essay on "Evaluating the Role of Digital Storytelling in Building Brand Identity for Higher Education Institutions: Opportunities and Challenges"

Copied

Source: Sarah Newman, metaLAB (at) Harvard

02 Human First, AI Review

Description

Students first complete a task independently, then use generative AI tools to review and improve their work. AI acts as a second-opinion reviewer, suggesting revisions, identifying gaps, or raising questions. Students critically assess the AI feedback, decide which suggestions to adopt or reject, and reflect on how AI input shaped their final outcome.

Example

Revising a Draft with AI Feedback

1. Students write an essay draft without using AI.

2. Students input their draft into Copilot or ChatGPT to request revision suggestions or feedback.

3. Students evaluate the AI-given feedback and suggestions, consulting with their instructors before editing based on the AI’s suggestions. Students copy their original draft onto a new page and apply edits, keeping the old version intact.

4. Students submit both old and revised drafts.

PROMPT:

Provide revision suggestions or feedback on the draft below

"In the competitive landscape of higher education...ultimately strengthening their brand in a competitive landscape." (Insert student's draft)

Source: Jun Wang, University of Virginia

03 AI-Generated Materials for Analysis

Description

AI produces sample materials for analysis, such as case studies, scenarios, or artefacts which students then apply their disciplinary analysis.

Example

Analysing AI Imitation of Literary Style

1. Students prompt a generative AI tool to write a passage imitating the style of a well-known author (e.g. Virginia Woolf or Cormac McCarthy).

2. They identify five hypotheses explaining specific stylistic choices made by the AI, referencing sentence structure, diction, tone, and grammar.

3. Students compare their observations to published analyses of the author’s style, and write a 500-word analysis discussing where the AI succeeded or failed in emulating it.

PROMPT:

Write a passage imitating the style of Virginia Woolf

Copied

Source: Chris Lott, University of Washington

04 AI for Immersive Learning

Description

Students use generative AI tools to create rich experiences that deepen their understanding of disciplinary concepts. By crafting simulations, environments, characters, or narratives, students are transported into the world of the content—whether historical, scientific, or conceptual. The goal is to promote deeper engagement, creative exploration, and personal connection to complex ideas using AI as a storytelling or visualisation partner.

Example

Immersive Historical Narrative

1. Students choose a major historical event (e.g. French Revolution) and select a framing perspective (e.g. adolescent artisan).

2. They input the prompt sample provided by instructors or of their own to AI tools to generate a detailed first-person narrative of life during that time. The narrative includes key figures, daily routines, sociopolitical dynamics, and sensory detail rooted in historical fact. Students also create an AI-generated image depicting themselves in the setting

PROMPT:

Generate a detailed first-person narrative of life from "adolescent artisan" during "French Revolution". The narrative includes key figures, daily routines, sociopolitical dynamics, and sensory detail rooted in historical fact.

Copied

PROMPT:

Generate image depicting the narrative.

Copied

Source: Tim Mousel, Lone Star College

05 AI-Guided Self-Assessment & Reflection

Description

Students engage in a structured conversation with an AI tool to test, explain, or evaluate their understanding of a concept or argument. The AI acts as a dialogic partner—asking probing questions, offering feedback, and prompting clarification. Students document the interaction, reflect on the AI’s feedback, and analyse how it influenced their thinking.

Example

Concept Explanation Using the Company Law

1. The instructor introduces the "Company Law" and provides a prompt where the AI plays the role of a beginner, asks probing questions, and summarises the student’s understanding at the end.

2. Students interact with the AI and explain the concept they studied. The AI asks one question at a time with around 10 questions total, challenges vague or unclear points, and provides a summary of strengths and gaps.

3. Students save and submit the full dialogue.

PROMPT:

Plays the role of a beginner in introducing "Company Law", asks probing questions, and summarises my understanding at the end.

Copied

Source: Jamie Jirout, University of Virginia

The Next Era of Assessment: