Idea to Learning Experience: Five Questions to Ask Before Sharing an AI-Built Tool with Students
Generative AI makes it easier for faculty to turn an idea into a digital learning experience. Faculty can develop quizzes, simulations, calculators, interactive case studies, and practice activities by writing a prompt, testing the result, and making adjustments.
These possibilities invite experimentation. Faculty can try new approaches, build activities around specific course content, and offer experiences that existing platforms may not provide.
Getting an activity to work is only one step toward making it ready for students. Before sharing an AI-built tool with students, it is also helpful to consider its learning purpose, accessibility, privacy, accuracy, and long-term use.
The following five questions can help faculty move an idea from a working prototype to a student-ready learning experience.
1. Does the tool support a clear learning goal?
Start with what you want students to learn or practice.
A custom tool can help students explore a concept, apply knowledge to a scenario, practice a skill, make decisions, or receive formative feedback. Connecting to a specific learning objective helps ensure that the technology adds value to the experience.
For example, an interactive scenario might ask students to examine information, choose a response, and reflect on the consequences of their decision. In that case, the tool supports application and critical thinking. If the tool simply produces an answer, it may not support the thinking students need to practice.
It may also be useful to consider whether a custom AI-built tool is the best fit. In some cases, a Canvas quiz, discussion, worksheet, or another VCU-supported tool may meet the same learning goal more simply.
Questions to consider:
- What should students know or be able to do after using the tool?
- What will students actively do while using it?
- Does the tool encourage practice, reflection, decision-making, or critical thinking?
- Does technology make the learning experience more meaningful or useful?
- Could an existing VCU-supported tool accomplish the same goal?
The best starting point is not “What can AI build?” but “What would help my students learn?”
2. Is it accessible and usable across devices?
Students access course materials in different ways. Some use screen readers, keyboard navigation, magnification software, smartphones, tablets, Chromebooks, or slower internet connections. Others may encounter temporary barriers, such as a broken mouse or an environment where they cannot listen to audio.
Testing an AI-built tool with different users, devices, and ways of navigating can uncover issues that may not be obvious during development.
Helpful areas to review include keyboard access, color contrast, heading structure, alternative text, captions, form labels, instructions, error messages, and mobile responsiveness. VCU’s accessibility requirements are based on Web Content Accessibility Guidelines 2.1 Level AA.
Accessibility also includes the practical experience of using the tool. Consider whether students need to create a new account, purchase access, provide personal information, or use a particular type of device.
Questions to consider:
- Can students complete the activity using only a keyboard?
- Are buttons, form fields, instructions, and error messages clearly labeled?
- Does the tool work with screen readers and other assistive technologies?
- Does it work on both mobile and desktop devices?
- Are captions, transcripts, or text alternatives available for multimedia?
- Does it require an external account, subscription, or specific device?
- Is there another way to complete the activity if a student cannot use the tool?
Building accessibility into the process from the beginning usually makes a tool easier for everyone to use.
3. Does it protect student privacy and data?
An AI-built tool may connect to other platforms or services behind the scenes. Before asking students to use it, take time to understand what information it collects, where that information goes, and whether it is stored.
This is especially important when a tool connects to an external AI model through an API. Even when the activity appears inside Canvas or on a webpage, student entries may be sent to another service for processing.
Whenever possible, design the activity so students do not need to enter names, grades, student identification numbers, personal details, or other sensitive information.
VCU provides enterprise versions of tools such as Google Gemini, Microsoft Copilot, and Adobe Firefly. According to the VCU Office of the CIO’s generative AI guidance, these enterprise licenses include protections for university data. Free, public, or personally licensed versions of the same tools may handle information differently.
If a new tool will collect, store, process, or transmit VCU information, faculty can consult VCU Technology Services or IT Governance for guidance before using it with students.
Questions to consider:
- What information does the tool ask students to enter?
- Does it send information to an external AI model, vendor, or API?
- Is the technology approved for use with VCU data?
- Can students use the tool without entering personal information?
- Are students told when they are interacting with AI?
- Is it clear how their information will be used?
- Could the activity work without collecting or saving student responses?
A simple approach is often the safest: collect only the information the activity truly needs.
4. Have its content, calculations, and feedback been tested?
Generative AI can produce polished code quickly, but the first version may not handle every response or situation correctly.
A tool might work well with the examples used during development but respond differently when a student enters an unexpected answer, skips a question, or interprets the instructions in another way. Calculations, links, branching paths, and automated feedback should all be reviewed before the activity is assigned.
Testing does not have to happen all at once. Faculty can start with a small group of colleagues, student workers, or other reviewers and ask them to use the tool in different ways. Subject-matter experts can review the accuracy of its content, while accessibility and technology specialists can help identify usability issues.
Questions to consider:
- Has someone with subject-matter expertise reviewed the content?
- Have calculations and answer paths been checked?
- What happens when a student enters an incomplete or unexpected response?
- Are instructions and feedback clear?
- Could any feedback unintentionally reinforce a misconception?
- Are explanations grounded in reliable sources?
- Has someone other than the tool’s creator tested it?
AI-generated feedback can be useful for practice and reflection, but faculty should continue to oversee grades and other decisions that significantly affect students. VCU’s responsible AI guidance also recommends checking AI-generated information against reliable sources.
A small test group can be a great way to gather feedback and improve the tool before using it with an entire class.
5. Can the tool be maintained if the course or technology changes?
Courses and technologies both change over time. Readings are updated, assignments are revised, and learning objectives evolve. AI platforms, browsers, APIs, and learning management systems also release updates that may affect how a tool works.
A little documentation can make future updates much easier. Keep copies of the prompts, code, instructions, settings, and source materials used to create the tool. It is also helpful to note which accounts or services it depends on and who can make changes.
Questions to consider:
- Who will maintain or update the tool?
- Are its prompts, code, settings, and source materials documented?
- Does it depend on a personal account, API key, subscription, or payment method?
- Could someone else update it if needed?
- How might changes to the course affect the tool?
- What support will students receive if they encounter a problem?
- Is there a backup activity available if the tool is temporarily unavailable?
Planning for maintenance does not mean anticipating every possible problem. It simply means making sure the learning experience can continue if the technology changes.
A Quick Pre-Launch Check
Before sharing an AI-built tool with students, confirm that:
- The tool supports a clear learning objective.
- Students actively practice, apply, evaluate, or reflect while using it.
- Students understand when and how AI is involved.
- The tool does not ask students to enter sensitive information into an unapproved system.
- Accessibility and mobile usability have been tested.
- Students do not need to purchase access or create an unnecessary account.
- Content, calculations, and feedback have been reviewed for accuracy.
- A faculty member remains involved in grading and important decisions.
- The prompts, code, settings, and source materials are documented.
- Support and an alternative activity are available if needed.
From an Idea to a Learning Experience
AI-assisted development gives faculty another way to experiment with creative approaches to teaching. It can help turn an idea into a prototype quickly and make custom learning experiences more approachable.
The next step is to look at the tool through the students’ eyes. Is its purpose clear? Can everyone use it? Does it handle information appropriately? Is its content dependable? Can it continue to support the course over time?
These questions are not meant to discourage experimentation. They provide a practical way to refine an idea, catch concerns early, and create a stronger learning experience.
AI can help build the tool. Thoughtful planning and testing help make it ready for students.
Faculty interested in creating or reviewing an AI-supported learning experience can contact LEDstudio for help with learning design, accessibility, testing, and implementation.
Resources
Instructure Community. (2025, February 1). Incorporating html/css and js code. Canvas LMS Question Forum. https://community.instructure.com/en/discussion/632864/incorporating-html-css-and-js-code
Instructure Community. How do I upload custom JavaScript and CSS files to an account? Canvas LMS Admin Guide. https://community.instructure.com/en/kb/articles/661411-how-do-i-upload-custom-javascript-and-css-files-to-an-account
The Human Movement. The 7 Principles. https://thehumanmovement.org/
UNESCO. (2021, November). Recommendation on the Ethics of Artificial Intelligence. Global AI Ethics and Governance Observatory.
Virginia Commonwealth University, IT Governance. VCU Data Classification Tool. Division of Administration. https://itgovernance.vcu.edu/security/data-classification/
Virginia Commonwealth University, Office of the CIO. Generative AI at VCU. Technology Services. https://cio.ts.vcu.edu/ai/