Ethical AI Integration in Instruction

This faculty-facing guide supports Vector Technology Institute (VTI) instructors in designing and managing responsible uses of generative AI in teaching, assignments, feedback, and assessment. Its focus is pedagogical: define the permitted use, protect people and information, verify outputs, and keep student learning visible. For standards, attribution, misconduct, evidence, and response, use the Academic Integrity Faculty Guide.

---

Start with Assignment-Level Permissions

There is no single AI rule that fits every learning outcome. Before an assignment begins, state what students may do, what they may not do, which parts of the work may use AI, what disclosure is required, and how students remain responsible for the final submission. Align the wording with the assessment purpose and applicable VTI requirements.

Level 0: No AI for this assessment

Use this level when the assessment is intended to measure unaided recall, in-class writing, personal reasoning, or an individual performance that would be changed by AI assistance. State the boundary before the assessment and explain which accessibility or support arrangements remain available.

Level 1: Limited support with disclosure

Permit defined support such as brainstorming, grammar and clarity feedback, practice questions, translation for comprehension, or code debugging. Require students to disclose the tool and purpose, check every output, and submit their own reasoning and final work.

Level 2: Guided collaboration

Permit students to use AI for a named task such as comparing explanations, generating test cases, proposing an outline, or receiving draft feedback. Assess the student analysis of the output, the revisions made, and the reasons for accepting or rejecting suggestions.

Level 3: AI as the object or method of study

Design the assessment around testing, critiquing, documenting, or improving an AI output. Students should identify the tool and prompt context, verify claims, discuss limitations, and demonstrate subject knowledge beyond reproducing the output.

Transparency and Disclosure

When AI use is permitted, provide a simple disclosure expectation. Ask students to name the tool, date or stage of use, purpose, and the parts of the work affected. A disclosure is not a substitute for quality or attribution, but it makes the learning process visible and gives faculty a basis for discussing responsible use.

I used [tool] on [date or stage] for [brainstorming, feedback, drafting, coding, or another purpose]. I checked the output against course sources and my own work, corrected errors, and remain responsible for the final analysis, citations, code, and submission.

Adapt this wording to the assignment. Do not require a disclosure statement that conflicts with a no-AI assessment or treats every tool use as equivalent.

Verification Is Part of the Assignment

• Check every factual claim against authoritative course or library sources.
• Open and verify each suggested citation, quotation, DOI, URL, statistic, and reference detail.
• Test generated code, calculations, formulas, and configuration steps in the relevant environment.
• Ask students to explain why a source, method, answer, or revision is appropriate.
• Assess whether the final work demonstrates the learning outcome rather than only polished language.

Privacy, Confidentiality, Copyright, and Attribution

• Do not paste identifiable student information, private feedback, unpublished assessments, client information, credentials, or confidential institutional data into an AI service without an approved basis and appropriate safeguards.
• Consider whether prompts or uploads may be retained, reused, or visible to a provider. Use the minimum information needed.
• Check the rights and license for source material, student work, images, code, and AI outputs before redistributing or publishing them.
• Require attribution or disclosure when AI materially contributes to an assignment, teaching material, image, code sample, or other work, according to the task instructions.
• Do not represent an AI-generated citation or quotation as a source that was actually read and verified.

Hallucinations, Bias, and Other Limitations

Generative AI can produce confident but false claims, invented references, incomplete explanations, biased examples, insecure code, and outputs that reflect limited or unbalanced training data. It can also obscure uncertainty and make a weak answer sound authoritative. Teach students to question the output, compare perspectives, check context, and record corrections rather than treating fluency as accuracy.

Appropriate Uses by Teaching Context

Brainstorming and planning

Appropriate: generate possible questions, compare angles, or identify terms for further research. The student should select, refine, and support the final direction with course sources.

Feedback and editing

Appropriate: request suggestions about clarity, structure, or grammar. The student should decide which changes to accept and preserve their own meaning, evidence, and voice.

Drafting

Use caution: AI may help model a structure or provide a comparison draft when permitted, but it should not replace the student argument, analysis, source reading, or required writing process.

Coding and technical work

Appropriate when defined: ask for explanations, debugging ideas, test cases, or alternative approaches. Require students to run tests, understand the code, document material assistance, and explain design choices.

Research assistance

Appropriate: brainstorm keywords, identify questions, or summarize a source that the student then reads and checks. Do not accept an AI-generated bibliography or literature summary without verification against the actual sources.

Assessment Design That Preserves Learning

• Use proposals, search plans, annotated bibliographies, outlines, drafts, and revision notes when they match the learning outcome.
• Include personal, local, current, practical, or comparative questions that require decisions and evidence.
• Ask for a short process reflection explaining sources, choices, revisions, and any AI assistance.
• Use in-class checkpoints, demonstrations, or brief oral explanations for selected high-value outcomes.
• Use a rubric that rewards reasoning, verification, source quality, process, and reflection instead of polished output alone.
• Keep access and equity in view: do not make advanced paid AI access a hidden requirement for success.

Sample Assignment Policy Language

No AI: AI tools may not be used for this assessment. The purpose is to evaluate unaided performance in [skill]. Ask the instructor before using any tool if you need an accommodation or clarification.

Limited use: You may use AI for [named tasks] only. Include a brief disclosure naming the tool, purpose, and affected section. You remain responsible for checking claims, citations, code, and the final submission.

Guided use: Use [tool or tool type] to produce and critique an output for [named task]. Submit the output or record, your verification notes, revisions, and a reflection on limitations. The assessed work is your analysis and decisions.

Responsible Faculty Use

Faculty may use AI to brainstorm examples, draft low-stakes explanations, generate practice questions, compare wording, or suggest feedback prompts. Review every output for accuracy, bias, accessibility, tone, copyright, and alignment with the course. Do not upload identifiable student work or confidential assessment information without an approved basis, and do not delegate grading, progression, or other high-stakes academic judgments to an AI system.

When AI Use Becomes an Integrity Question

If a student’s use appears to conflict with the assignment instructions, move from teaching guidance to a careful evidence-based review. Check the stated permission, disclosure, source and process evidence, and student explanation, then follow the applicable VTI process. The Academic Integrity Faculty Guide covers misconduct types, attribution, documentation, fair response, and student support.

Which Guide Should I Use?

Use Academic Integrity Faculty Guide for standards, attribution, misconduct, evidence, documentation, and response.

Use this Ethical AI Integration in Instruction guide for defining acceptable AI use, disclosure, verification, privacy, assessment design, and responsible faculty use of AI in teaching and assignments.

For source evaluation and citation support, see the Research Skills Guide and APA Formatting and Citation (7th Edition).