Generative AI tools can write essays, solve problems, and answer test questions. More recently, agentic browsers and assistants can act on their own inside a web page. That means a tool can, in some cases, work through an online quiz or draft an assignment with little input from the student.

This affects how much a piece of submitted work tells you about what a student knows. It does not mean your assessments no longer work. It means the design of an assessment, and the setting in which students complete it, matter more than they used to. The rest of this page offers ways to respond that keep your learning goals at the center.

Before you change any assignments, decide what role AI should play in your course. You are the one who sets this, and reasonable instructors land in different places depending on the field and the outcomes they are teaching.

A range of choices

• Set AI aside for certain work. Some skills need to be practiced without help, so you may ask students to complete specific work on their own.
• Allow AI with conditions. You may permit AI for some steps, such as brainstorming or checking grammar, and ask students to say how they used it.
• Build AI into the work. You may treat AI use as a skill to teach, and design tasks where students use it and evaluate what it produces.

Authentic assessments ask students to do something close to real work in your field, using their own experience and judgment. These tasks are harder for a tool to complete well, and they tend to be more meaningful for students. Here are practical ways to move in that direction.

• Ask for the process, not only the product. Collect an outline, a draft, notes, or a short log of decisions along with the final work.
• Ground the task in your course. Ask students to use a specific reading, a class discussion, a data set from lab, or an example from a recent session.
• Connect it to the student. Ask students to apply an idea to their own experience, their community, or a local case.
• Ask for application and analysis. Move from recall toward tasks where students weigh options, justify a choice, or solve a problem specific to a situation you provide.
• Add a spoken element. Ask students to explain their work in a short recording or a brief conversation, so they show their thinking.
• Build in reflection. Ask students to describe what was hard, what they changed, and why.
• Use stages. Break a large project into steps with feedback, so the work develops over time and you see it grow.

You do not have to redesign a whole course at once. Changing one or two key assessments a term is a reasonable pace, and OIR can help you choose where to start.

Some assessments need to be taken independently, such as a licensure-style exam, a gateway test, or a high-stakes final. For these, Honorlock gives you a way to offer a secure online test without moving it to a classroom. Honorlock is Wichita State's proctoring tool, and WSU covers the cost.

How Honorlock helps with AI during a test

• It proctors the testing session with automated monitoring, so students take the test in a controlled setting rather than an open="" browser.
• It records the session and flags activity that looks unusual for you to review, which supports your judgment about whether outside tools were used.
• It applies within the testing window you set, so students keep the flexibility to test when they are able.

Learn how to set it up on the WSU Honorlock pages:

• Honorlock for Instructors
• Honorlock in Blackboard
• Honorlock Proctoring overview

Moving some work into the room is one option, and it can be the right call for a few key moments. You do not have to convert an entire course to use it.

• Hold a short in-class writing sample early in the term, so you have a sense of each student's own voice.
• Use a brief oral check or presentation for a major project, so students explain their work in person.
• Give a proctored in-class exam for the outcomes that most need independent demonstration.

For fully online courses, Honorlock and the authentic design ideas above cover much of the same ground without requiring a campus visit.

If your policy allows AI, you can make its use part of the learning. Students in many fields will use these tools in their work, and there is value in teaching them to use them with judgment.

• Ask students to generate a draft with AI, then critique and correct it against course standards.
• Ask students to compare their own answer with an AI answer and explain the differences.
• Ask students to document their prompts and describe what they accepted, changed, or rejected.
• Discuss the limits of these tools, including errors, bias, and gaps, so students learn to check the output.

This approach pairs well with authentic design, since both ask students to show their own thinking about the work.

1. Write your AI policy and add it to your syllabus.
2. Pick the one or two assessments that matter most, and decide how each should be handled: redesign it to be more authentic, allow AI with conditions, or secure it with Honorlock or an in-person option.
3. Make the changes to those assessments, and state the expectations on each one.
4. Review how it went at the end of the term, and adjust for the next one.