How Do I Build Student AI Literacy and Fluency?

Generative AI is increasingly becoming part of how students learn, work, and access information. Preparing students to use these technologies effectively requires more than teaching them how to write prompts or establishing rules about when AI is permitted. Students need opportunities to develop the knowledge, judgment, and experience necessary to use AI purposefully, critically, and responsibly.

AI literacy involves understanding how AI tools work at a practical level, recognizing their capabilities and limitations, evaluating their outputs, and understanding the ethical and professional considerations associated with their use.

AI fluency goes a step further. It involves knowing when, why, and how to use AI effectively, how to adapt its use to a particular task, and when another approach is more appropriate.

These abilities develop through practice. Rather than treating AI literacy as a one-time topic, look for small opportunities throughout your course for students to use, evaluate, critique, and make decisions about AI in the context of what they are already learning.

Make Your Expectations for AI Explicit

Students encounter very different expectations for AI across courses, instructors, and assignments. Do not assume that students know what constitutes appropriate use in your course.

Clearly communicate:

• When AI use is permitted, encouraged, restricted, or prohibited
• What types of AI assistance are appropriate for different activities
• Whether and how students should disclose or acknowledge AI use
• What students remain responsible for when AI is used
• Why particular boundaries exist

Explaining the why behind your expectations helps students develop judgment they can transfer beyond a single assignment.

Let Students Examine What AI Does Well and Poorly

One of the simplest ways to build AI literacy is to make AI output an object of analysis. Provide students with an AI-generated response and ask them to evaluate it. For example:

• What did the AI get right?
• What is inaccurate, incomplete, or misleading?
• What important information or nuance is missing?
• What assumptions did the response make?
• What would need to change before you would trust or use this response?
• How would you verify its claims?

The goal is not simply to demonstrate that AI makes mistakes. Students should learn that AI output requires active evaluation rather than passive acceptance. As students develop more expertise, increase the complexity of the AI outputs they evaluate. Obvious errors require little judgment. Plausible but incomplete or subtly flawed responses provide richer opportunities for developing AI literacy.

Ask Students to Improve AI Output

Critiquing AI can be taken one step further by asking students to revise what it produces. Have students generate or receive an AI response and then ask them to correct, refine, expand, reorganize, or otherwise improve it. This changes the student’s role from consumer to editor. 

To improve an AI-generated response successfully, students must draw upon their own knowledge and determine what a high-quality response should contain. Consider asking students to explain the changes they made and why. Their rationale may provide more evidence of learning than the revised product itself.

Teach Students to Use AI Iteratively

Effective AI use rarely consists of entering one prompt and accepting the first response. Teach students to  refine prompts, provide additional context, ask follow-up questions, challenge an initial response, request alternative approaches, and evaluate whether subsequent outputs actually improve.

Students can also compare how different instructions change the quality or usefulness of a response. The goal is not to teach a collection of “perfect prompts.” It is to help students understand that working effectively with AI is an iterative process involving communication, evaluation, and refinement.

Require Verification

Students should learn that using AI does not transfer responsibility for accuracy to the technology. When students use AI for academic work, periodically require them to verify important claims using appropriate sources or methods. Ask them to identify what they verified, how they verified it, and whether the verification changed their assessment of the AI response.

This reinforces an important habit: AI can assist with a task, but the user remains responsible for determining whether its output is trustworthy and appropriate.

Ask Students to Decide Whether AI Should Be Used

AI fluency includes knowing when not to use AI. Present students with different tasks (learning activities, professional procedures, etc.) and ask them to decide whether AI would be appropriate or useful. They should be able to explain their reasoning.

Possible considerations include:

• Does AI meaningfully improve the task?
• What could be gained by using it?
• What could be lost?
• What expertise would someone need to evaluate the output?
• What are the consequences if the output is wrong?
• Are there privacy, ethical, professional, or other concerns?
• Would using AI undermine the learning the activity is intended to produce?

This moves students beyond a simple permitted/prohibited model and toward context-dependent professional judgment.

Make Thinking Visible

When students use AI, consider assessing not only the final product but also the thinking surrounding its use. Depending on the activity, students might briefly document:

• Why they chose to use AI
• What they asked it to do
• How they evaluated the response
• What they accepted, rejected, or changed
• How they verified important information
• What they ultimately contributed to the work themselves

These do not need to become lengthy AI-use reports. Even a brief reflection can shift attention from “Did you use AI?” to the more educationally useful question: “How did you exercise judgment while using it?”