Lamar University AI Policy

Faculty should be clear with students about course-specific policies on permitted AI use, if any, and students are encouraged to ask instructors directly when a policy is unclear. See the syllabi guidance later in this document for sample language.

Sample language: “Generative AI tool use is strictly prohibited in this course. All work submitted,

including drafts and brainstorming artifacts, is assumed to be independently produced by the student.

Use of generative AI tools for any part of a graded assignment will be treated as a violation of academic

honesty; see the student handbook for possible consequences. Questions about what constitutes a

violation should be directed to the instructor as soon as they arise.”

Sample language: “Generative AI tools are allowed in specified contexts in this course, with attribution

required. Students must credit AI tools whenever used, including for idea generation, and include an

• In-Class Prohibition: AI tools are not permitted during in-class

examinations or assignments unless explicitly allowed and instructed.

Instructors shall take reasonable and necessary precautions, including the careful administration and monitoring of examinations and assignments, to prevent acts of academic dishonesty.

Lamar University encourages responsible exploration of generative AI, balancing innovation

with the guidelines above.

• The course recognizes the potential of generative AI tools for brainstorming,

exploring responses, and creative engagement with materials.

• Potential uses. Content generation, course development assistance, brainstorming, data analysis, document summarization, learning new skills, writing code, and more. Faculty should visit CTLE for ideas.

Generative Artificial Intelligence (AI) is a form of artificial intelligence trained on large datasets to

produce content such as text, images, music, video, and code in response to user input.

Faculty should be clear with students about course-specific policies on permitted AI use, if any,

and students are encouraged to ask instructors directly when a policy is unclear. See the syllabi

guidance later in this document for sample language.

• Potential uses. Content generation, course development assistance, brainstorming, data analysis, document summarization, learning new skills, writing code, and more. Faculty should visit CTLE for ideas.

Generative AI cannot fulfill core authorship responsibilities such as approving the final version,

accounting for the accuracy and integrity of the work, disclosing conflicts of interest, or managing

copyright and licensing obligations. Accordingly, AI tools should not be listed as authors on scholarly

works (COPE, 2023; Zielinski et al., 2023; Flanagin et al., 2023; Nature, 2023).

Authors using AI tools in writing, analysis, or graphical elements must disclose that use in the

Materials and Methods section (or equivalent). Authors remain responsible for the accuracy of

AI-assisted output and should review and edit carefully, since AI can generate authoritative-sounding but inaccurate or biased content (ICMJE, 2023; Hosseini, Rasmussen & Resnik,

2023).

Authors using AI tools in writing, analysis, or graphical elements must disclose that use in the

Materials and Methods section (or equivalent).

Non-public or IRB-governed research data is Sensitive, Confidential, or Regulated under

Lamar's data classification framework and may not be entered into any AI tool, including

Copilot, without prior review. De-identification alone does not authorize AI use of IRB-regulated

data; re-identification risk and IRB protocol terms still apply.

Researchers seeking to use AI tools with non-public research data must complete a DARpA AI

Data Use Request Form prior to that use.

The same concerns that apply to manuscripts extend to grant applications. Funding agencies

hold applicants responsible for any AI-introduced plagiarism, falsification, or fabrication in

submitted materials (Lauer, Constant, & Wernimont, 2023).

Non-public or IRB-governed research data is Sensitive, Confidential, or Regulated under

Lamar's data classification framework and may not be entered into any AI tool, including

Copilot, without prior review. De-identification alone does not authorize AI use of IRB-regulated

data; re-identification risk and IRB protocol terms still apply.

Sample language: “Generative AI tools are allowed in specified contexts in this course, with attribution

required. Students must credit AI tools whenever used, including for idea generation, and include an

Instructors and students should reflect on how to properly acknowledge and cite content derived

from generative AI tools, including its role in brainstorming and argument development.

• Students must give credit to AI tools when used, even if only for generating

ideas.

• Appendix Requirement: Assignments involving AI tools must include an

appendix highlighting the relevant sections, specifying the AI tools used,

explaining how they were used, and providing reasons for their use.

Authors using AI tools in writing, analysis, or graphical elements must disclose that use in the

Materials and Methods section (or equivalent).

AI detection tools are not approved for use in evaluating student work at Lamar University. Their

output is not a reliable basis for an academic integrity finding, and their use raises the same

student data and consent concerns as any other AI tool. Concerns about the originality of

student work should be addressed through existing academic integrity procedures, not through

AI detection software.

Use of generative AI tools for any part of a graded assignment will be treated as a violation of academic

honesty; see the student handbook for possible consequences.

Procedures for discipline due to academic dishonesty shall be the same as in other violations of the Student Code of Conduct , except that all academic dishonesty cases shall be considered and reviewed by the faculty member, and if necessary, the Department Chair, Dean, and Provost. The faculty member shall conduct a complete, thorough, and impartial investigation of the charge of academic dishonesty and determine whether or not the student was responsible for the violation.

AI tools must not be used to generate non-public content, including proprietary or unpublished

research, legal analysis, recruitment decisions, completion of academic work not authorized by

the instructor, non-public instructional materials, or direct grading of student work.

Student submissions are both the intellectual property of the student and protected educational

records under the Family Educational Rights and Privacy Act (FERPA). Faculty may not upload

student work to AI platforms, including for grading, feedback generation, or AI-generated-content detection, without the student's explicit consent.

Faculty should be clear with students about course-specific policies on permitted AI use, if any,

and students are encouraged to ask instructors directly when a policy is unclear.

Microsoft Copilot is Lamar University's approved generative AI platform, in accordance with the

university's Generative Artificial Intelligence policy. Microsoft 365 systems, including Copilot,

have undergone university compliance review.

Copilot's interface displays a message indicating that personal and company data are protected

within that chat. This does not mean any information may be entered freely. The data

classification guidance below still applies regardless of platform.

• Public — information intended for unrestricted release (e.g., published web content, job

postings, campus maps, university directory information). Freely usable in any AI tool.

• Sensitive — information not intended for public release but with limited risk if disclosed

(e.g., internal memos, non-public policy drafts, unpublished research data at the data

owner's discretion). Use only within Copilot's protected institutional environment, and

only as necessary.

• Confidential — information that could cause material harm if disclosed (e.g., employee

performance records, salary information, legal affairs and institutional relations matters,

contracts). Must not be entered into any AI tool, including Copilot, absent a specific data

agreement covering that use.

• Regulated — information subject to legal or regulatory protection (e.g., Social Security

numbers, financial account numbers, biometric identifiers, health information including

PHI, export-controlled information). Must never be entered into any AI tool.

AI tools must not be used with Confidential or Regulated data as defined above, unless a

university contract specifically protects that data within the tool in question.

Lamar University supports responsible exploration of generative AI tools. Users should remain mindful

of information security, data privacy, compliance standards, copyright, and academic integrity

when using these tools.

AI in Every Program (AIEP) note: This guidance complements, and should be read alongside, Lamar's

formal Generative Artificial Intelligence policy and the AIEP framework maintained by University

Innovation. Where this document and the formal policy differ, the formal policy governs.

Lamar University encourages responsible exploration of generative AI, balancing innovation

with the guidelines above.

• Review this guidance and the university's formal AI policy. Confirm compliance with data

classification requirements before using any confidential or regulated information.

• Check which tools are currently approved and their data handling levels. Some colleges

may offer additional tools for eligible users.