One instructor uses AI every week. Another avoids it because they do not trust the answers. A third wants to try it but is unsure what information can be uploaded. A single presentation about prompt writing will not resolve those differences.
AI literacy for instructors means being able to choose an appropriate use, understand the tool's limits, verify its output, protect information, and explain the rules to students. It is a teaching and judgment skill as much as a technical skill.
For a training school, the goal is a shared standard of practice. Staff do not need identical enthusiasm for every new tool. They do need to know how to use approved tools responsibly and when to rely on another approach.
Build confidence around five practical capabilities
UNESCO's AI Competency Framework for Teachers organizes 15 competencies across five dimensions, including human-centered thinking, ethics, AI foundations, pedagogy, and professional learning. It is a useful reference for planning staff development.
The following five capabilities are a practical Steams Online training exercise, not a reproduction of UNESCO's framework or a formal certification.
1. Explain what the tool is doing
An instructor should understand that fluent generated text can still be wrong. The tool may omit a condition, mix sources, invent a citation, or answer a different question from the one intended.
Ask staff to describe an AI response as a draft to inspect rather than a source to trust automatically. They should also distinguish the tool's response from the underlying reference that supports a claim.
2. Choose a task with a clear purpose
Good starting tasks are bounded and easy to check: drafting an orientation outline from approved notes, proposing practice questions, or rewriting a school-created explanation for clarity.
Tasks involving grades, student conduct, clinical decisions, or sensitive records need additional controls and may be unsuitable for a particular tool. The instructor should be able to explain why AI assistance is useful for the task—not simply that it is available.
3. Verify before using
Give staff an intentionally imperfect sample. Include one unsupported claim, one ambiguous quiz answer, and one citation that does not support the paragraph.
Have them identify the problems, locate an approved source, and rewrite the material. The exercise should assess their review process, not how quickly they can produce a prompt.
4. Protect information
Use fictional cases during training. Ask instructors to identify information that should not go into an unapproved tool: student identifiers, patient details, login credentials, payment information, or confidential documents.
Then show the school's approved-tool list and reporting process. Staff need a practical answer to “Who do I ask?” The U.S. Department of Education's privacy and education technology resources can support a school's review of how educational services handle information.
5. Teach students the boundaries
An instructor should be able to write an assignment instruction that identifies permitted assistance, prohibited assistance, disclosure, and how learning will be assessed.
For example: “You may use an approved AI assistant to generate practice questions for studying. The submitted reflection must be your own analysis. Disclose any permitted editing assistance and verify factual statements against the assigned material.”
That example needs adaptation to the assignment. It illustrates the level of specificity students should receive.
A 90-minute workshop your school can adapt
This is an example agenda, not a prescribed training requirement.
Minutes 0–15: Start with a school task. Ask participants to name one repetitive preparation task and one decision they would not delegate to AI. Discuss the reasons.
Minutes 15–30: Compare two responses. Use the same approved lesson excerpt with a broad prompt and a carefully bounded prompt. Discuss what improved and what still needs checking.
Minutes 30–50: Find the errors. In pairs, review a prepared output containing factual and instructional problems. Ask each pair to explain how they verified its claims.
Minutes 50–65: Apply the privacy rules. Sort fictional examples into approved, requires review, and prohibited uses under your school's policy. Resolve disagreements explicitly.
Minutes 65–80: Write student instructions. Have each instructor revise one real assignment's AI-use guidance. A colleague should be able to understand the boundaries without additional explanation.
Minutes 80–90: Demonstrate readiness. Each participant submits a corrected output, its supporting source, and a short explanation of the review decisions.
The facilitator should keep examples small enough to inspect fully. A short exercise done carefully is more useful than a large batch of generated material nobody has time to verify.
Use a simple readiness check
Before staff use AI-generated material in a course, ask them to demonstrate that they can:
- Identify the learning objective and approved source.
- Write a prompt that limits the task and information used.
- Locate and correct an unsupported statement.
- Explain what student information must remain private.
- Identify when human or specialist review is required.
- Write clear student-facing instructions about allowed use.
Record what was demonstrated and where more practice is needed. Attendance at a workshop is different from demonstrated competence in the workflow.
For an instructor who is struggling, offer a smaller starting task and a colleague review. For an experienced user, focus on consistent evidence, source quality, and documenting decisions rather than merely generating more content.
Support the work after the workshop
Maintain a short collection of approved prompts and reviewed examples. Include the purpose, appropriate inputs, limitations, and review owner. Retire examples when policies or source material change.
Hold a brief staff discussion after the first few weeks. Ask what saved useful effort, what required more checking than expected, and what confused students. Measure the quality of the result as well as time spent.
Do not reward output volume alone. A hundred weak questions are less useful than ten well-reviewed questions tied to the actual course.
Frequently asked questions
Do instructors need programming skills?
Usually not for everyday drafting and review tasks. They need enough understanding to evaluate outputs, handle information appropriately, and use the school's approved workflow.
Should every instructor use the same AI tool?
A small approved set can make training and oversight easier. Match tools to tasks and account requirements; do not assume an individual's free account has the same protections as an institutional service.
Can AI literacy training replace subject expertise?
No. Subject expertise is what lets instructors recognize missing context, incorrect answers, and unsafe suggestions. AI training should strengthen that judgment.
Turn staff development into a repeatable course process
Steams Online's LMS supports materials, assignments, quizzes, and progress tracking. Those are useful building blocks when organizing instructor development and collecting completed practice work.
Explore Steams Online with your team using a short staff-training module. Start with the evidence you want instructors to produce, then build the course around it.
