Student Use of Generative AI: Use, Disclosure, and Accountability

Governing Framework: WADL Assignment Authorization

Student use of generative AI is governed by this manual and the Wandt AI Disclosure Label (WADL) authorization stated for each assignment. WADL separates three questions that must not be collapsed into a single general permission:

• the permitted AI-assisted process level (P);
• the required verification level (V); and
• the required record and provenance level (R).

Each assignment establishes its permitted WADL process range and its required verification and record levels. Those assignment terms implement authority delegated by this manual; they do not override the manual. Students must follow the WADL version and instructions identified in the assignment and may consult the current public framework at usewadl.org.

AI use outside the assignment’s authorized process range, below its required verification level, or without its required records is unauthorized. If a student is uncertain how an assignment’s WADL authorization applies, the student must ask through Discord before using the tool.

Examples of generative AI tools include ChatGPT, Claude, Gemini, Copilot, Perplexity, and comparable text, image, audio, video, coding, research, or agentic systems. Built-in AI features count as AI use even when they appear inside a word processor, search engine, coding environment, note-taking tool, or other software.

Permitted Uses Within the Assigned WADL Range

Depending on the process level authorized for the assignment, permitted uses may include:

• brainstorming, outlining, and issue spotting;
• assistance in drafting or restructuring material;
• editing for clarity, organization, tone, grammar, or concision;
• study aids and self-quizzing;
• alternative arguments, counterarguments, or analytical structures;
• coding or technical support;
• accessibility support; and
• potential sources, authorities, search terms, or research directions used only as unverified leads.

The assignment’s WADL authorization determines which of these uses are permitted and how far the student may rely on them. AI may support a student’s process, but it does not reduce the student’s obligation to provide original analysis, accurate reasoning, verified claims, and work responsive to the actual assignment. Simply placing the assignment into an AI system and submitting the result does not demonstrate the required learning.

No-AI Assessments

AI tools are prohibited on a midterm, final examination, quiz, practical exercise, oral assessment, or other activity designated “No AI” or assigned WADL process level P0. This restriction includes using AI to generate, check, revise, translate, summarize, search for, or otherwise assist with an answer during the restricted assessment or assessment window.

When an assessment is marked “No AI,” students must disable or avoid built-in AI features and must not seek AI assistance through another device, account, person, or service. Any AI use on a restricted assessment is a serious academic-integrity violation.

Core Accountability: The Student Owns the Outcome

AI is a tool. Everything submitted under a student’s name is treated as that student’s work product, reasoning, representation, and claim.

The student is responsible for:

• every fact, quotation, citation, authority, calculation, interpretation, and conclusion;
• compliance with the assignment and rubric;
• the legality and ethics of the process used;
• identifying and correcting inaccurate, incomplete, biased, misleading, or fabricated output; and
• being able to explain the final submission and how it was produced.

Blaming a tool does not excuse an error or violation. Reckless reliance on AI, especially when it produces invented or misleading material, may be treated as an academic-integrity violation.

Required WADL Records and Exact Audit Trails

Each assignment’s WADL record level establishes what provenance material the student must create, retain, and submit. When the assigned record level requires prompts, outputs, verification journals, correction records, revision histories, attachments, generated files, or other workflow evidence, the records must be exact, complete, and unaltered.

A summary, paraphrase, reconstruction, approximation, or later-created substitute is not an exact record. Students are responsible for preserving the original records produced by the tool or workflow, including the complete session history required by the assignment. Required records must be organized so that they can be produced promptly upon request.

WADL-required records may include:

Prompt and Output Records

• the exact prompts or instructions given to the tool;
• the full outputs received, including relevant attachments or generated files;
• follow-up prompts and outputs that changed the result;
• the tool name and model or version; and
• the dates and times associated with the interaction.

Verification Records

When the assignment’s verification or record level requires a verification journal, it must accurately identify:

• what was checked;
• where it was checked, with primary or official sources strongly preferred;
• what was confirmed;
• what was corrected, qualified, or removed; and
• any source lead that could not be verified and therefore was not used.

Integration and Revision Records

When required by the assigned WADL level, the student must retain evidence showing how AI output affected the work and how the final submission differs. Examples include replacing an AI-suggested authority with controlling authority, correcting statutory elements, rewriting generated analysis, testing code, removing unsupported statistics, or preserving before-and-after text for editing use.

All records required by the assignment must be preserved until the final course grade is issued and any related review, appeal, or academic-integrity process is complete. Failure or refusal to create, retain, submit, or produce the exact records required by the assignment’s WADL level is itself a violation of this policy.

Required WADL Disclosure

Every submission must contain the WADL label or disclosure block required by the assignment, including a no-AI declaration when the assignment requires a label from every student. The label must accurately report the student’s actual process, verification, and record levels and must appear in the location and format stated in the assignment.

The disclosure label does not replace the underlying records, verification journal, prompt logs, attachments, or other evidence required by the assigned WADL level. A vague, incomplete, inaccurate, or misleading label does not satisfy the disclosure requirement. When uncertain between two process categories, students must disclose the higher applicable category.

Failing to include the required WADL label, understating AI involvement, claiming a verification level not actually completed, claiming a record level without maintaining the required records, or otherwise misrepresenting the workflow is a policy violation even if the resulting work is otherwise accurate.

Verification and No-Fabrication Standard

AI output is not authority. Students must complete the verification level assigned under WADL and independently verify all material required by that level. No WADL authorization permits fabrication or presentation of unverified AI output as established fact. Legal authorities should be checked against primary sources whenever possible. Citations must be real and must support the proposition stated.

Students may not submit:

• invented or altered authorities, quotations, holdings, facts, data, sources, or credentials;
• citations that do not exist or do not support the claim;
• an authority described as controlling when it is not;
• untested code or technical instructions presented as verified when meaningful testing was possible; or
• an AI-generated assertion presented as established fact without independent support.

Privacy, Confidentiality, and Course Materials

Students may not enter sensitive personal, protected, or confidential information into an AI system. This includes student identification numbers, grades, private communications, disability or medical information, disciplinary history, identifying information about classmates, private credentials, confidential workplace information, protected investigative data, and nonpublic personal data.

Students may provide an AI system with ordinary assignment instructions, prompts, and rubrics when doing so is within the assignment’s WADL-authorized process range and the material is not marked restricted, confidential, or nonpublic.

Students may not paste, upload, or expose protected course materials—including examinations, restricted or nonpublic scenarios, lectures, recordings, answer keys, peer work, or private Discord content—to an external AI system unless Professor Wandt expressly authorizes the specific material and use. WADL authorization does not by itself authorize disclosure of another person’s work, restricted assessment content, private communications, or sensitive information.

Students may use their own notes and their own work with an authorized AI tool, provided doing so does not disclose protected material, confidential information, or another person’s work.

Instructor Review and Audit Rights

To evaluate learning and protect academic integrity, Professor Wandt may:

• require the complete audit trail and exact records assigned under WADL before or after grading;
• ask the student to explain and defend the reasoning or process orally or in writing;
• compare the submission with drafts, version histories, citations, prompts, outputs, or source materials;
• require reproduction of part of the work under controlled conditions; and
• conduct authorship, source, technical, and integrity checks consistent with applicable policy.

Enforcement

Violations are treated seriously. Subject to the syllabus and applicable College and CUNY procedures, a violation may result in no credit or a failing grade on the affected work, a course-grade consequence up to course failure, and referral through the academic-integrity or disciplinary process.

Using AI on a “No AI” or P0 assessment, exceeding the WADL-authorized process level, failing to complete the required verification level, submitting fabricated authorities or evidence, concealing material AI use, or failing or refusing to produce the exact records required by the assigned WADL record level will ordinarily be treated as a major violation. The seriousness, intent, effect, and applicable institutional requirements will be considered in determining the course response and any referral.