Arizona State University · Office of the Provost + EdPlus

Creating Assessments in the Age of AI

A collaborative effort to help faculty cultivate and assess authentic student learning at a transformational moment. Explore the frameworks and strategies below, then tell us what you want to redesign.

What this is

Good course design is our strongest move

As AI becomes part of how students learn and work, many faculty are rethinking how their assessments can make student learning and thinking visible. These nine practical strategies are a starting point for strengthening assessments, clarifying expectations for AI use, and designing work worth doing. The Office of the Provost and EdPlus developed them through faculty workshops, listening sessions, research, and ongoing conversations about assessment in the age of AI, and the effort will keep growing and taking shape based on faculty input.

The strategies and frameworks on this site emerged from work we have been doing to help faculty navigate the challenges AI has brought to teaching and assessment. They shaped what we ran in July: the Assessment Design Accelerator, a two-day event we hosted with faculty invited from the top 100 enrolling ASU Online courses. About 25 participants joined us this summer, and it is something we will do more of going forward if there is faculty interest.

Along the way, we have tried to do the researching, reading, and testing so faculty do not have to start from scratch, surfacing the practical strategies and examples that have most helped us think through assessment design in this moment. A through-line to hold in mind: cheating prevention is a function of design, not solely of enforcement. These strategies redesign assignments and assessments so that authentic student learning becomes easier to see, including at scale, and so students are empowered to create work worth doing regardless of what AI can produce.

Our conviction runs through all of it: good course design is our strongest move in this environment. When we invest in it, we are in a much better position to meet the challenges AI has surfaced.

Two anchor frameworks

Two frameworks that anchor the work

They answer different questions. Held together, they turn abstract redesign into a single, concrete design sentence for every assignment.

TILT

Transparency in Learning and Teaching

A small, high-leverage change to how assignments are written. Many students, especially those first in their family to attend college, struggle not because the work is too hard but because the unwritten rules are unclear. TILT makes three things explicit on every assignment:

  • Purpose. Why am I doing this? The skills practiced, knowledge gained, and connection to course outcomes.
  • Task. What exactly do I do? The steps, what to avoid, and any productive-struggle left open on purpose.
  • Criteria. How is it judged? A checklist or rubric with annotated examples of successful work.

An AAC&U pilot across 7 minority-serving institutions found that TILTing just two assignments produced significant gains in confidence, belonging, and skill mastery, with the largest gains for first-generation and underserved students. It is transparency, not more work.

Read the full framework →
Two-Lane

Thinking about two purposes for assessment

From the University of Sydney (Liu & Bridgeman). Assessment now has to both assure that students achieved the outcomes and help them learn to work with AI. One assessment type cannot do both, so every assessment is sorted into one lane:

  • Lane 1: Controlled. Conditions controlled enough to give a defensible signal of the student’s own capability. This is about conditions, not location, so it fits async online. AI may be banned, restricted, or deliberately permitted in specified ways, depending on the learning outcome being assured.
  • Lane 2: Open. AI use is permitted and perhaps encouraged. Students demonstrate their skills and knowledge in the environment where they will actually use them, while also developing AI literacy: evaluating, refining, and taking responsibility for AI output rather than simply submitting it.

Assign lanes assessment by assessment. Put controlled conditions only where an outcome must be assured, and let most of the course live in the open lane. A single task can combine both: an open project capped by a short controlled check.

Read the full framework →
Putting them
together

Two-Lane decides which lane a task lives in, controlled (Lane 1) or open (Lane 2). TILT makes the rules of that lane explicit, especially in the open lane. With those decisions in place, the nine strategies below offer practical ways to redesign individual assessments.

Practical moves

Nine practical strategies

Each one comes from real faculty who have published materials, working models, and examples you might consider adapting for your course and the ASU environment. Each is meant to spur your thinking and lay groundwork for discussion with colleagues, instructional designers, and leadership. Select any card to open the full strategy: why it works, how to actually do it, and where to look.

Explicit expectations Presence & relationship Observed cognition Task design that resists AI substitution Process & feedback Lane 1 controlled assurance
1

Set an explicit AI-use policy for each assignment

Per-task guidance (a level on the AI Assessment Scale, or a red / yellow / green tag) beats one blanket syllabus rule.

Explicit expectations Read the strategy →
2

Open with short, unscripted student intros

A 60–90 second video or audio intro establishes each student’s voice and presence for the whole course.

Presence & relationship Read the strategy →
3

Build in brief interaction each module

A short synchronous or semi-synchronous touchpoint builds relationship and presence, and keeps students engaged.

Presence & relationship Read the strategy →
4

Embed brief low-stakes oral explanations

A one- to two-minute “micro-oral” where students explain their reasoning puts the focus on how they think, and gives you authentic evidence in their own voice.

Observed cognition Read the strategy →
5

Require local, personal, or course-specific context

A real workplace, community, or lived case deepens learning and is hard to outsource to AI.

Task design vs. AI substitution Read the strategy →
6

Reference students’ own prior work

Build a mid-term on a student’s Week 3 submission or a named classmate’s post, so course-internal context is required.

Task design vs. AI substitution Read the strategy →
7

Make work-in-progress and feedback part of the deliverable

Drafts, decision logs, and a response memo make the work something the student has to actually do.

Process & feedback Read the strategy →
8

Design for a real audience beyond the instructor

When a real person needs the work, “good” becomes concrete and the assignment justifies itself.

Task design vs. AI substitution Read the strategy →
9

Include a controlled (Lane 1) assessment

A controlled assessment gives you confidence the work reflects what the student knows and can do. Controlled, not AI-banned, not in-person.

Lane 1 controlled assurance Read the strategy →
Open the full nine strategies, with the how-to and sources →

The complete resource: why each move works, how to actually do it, and where to look, plus the two foundational frameworks in depth.

Companion resource

Integrity Through Design

Academic integrity is not just upheld, it is actively cultivated through thoughtful course design. Grounded in the values of Principled Innovation, this guide treats integrity as a design challenge that centers human dignity and a culture of shared responsibility and trust.

Integrity Through Design · Teaching in the Age of AI Open in Google Drive →

Tell us what you’re working on

Express your interest

Want to redesign an assessment with us, or just stay in the loop as this effort grows? Share a few details and we will follow up.