Teaching & Learning
- University-wide guidance says AI use for assignments is not uniform and depends on the instructor or program
- The undergraduate academic integrity page says some instructors may encourage use for specific assignments while others may prohibit it, and students should ask if unsure
- Graduate guidance says programs and faculty may set different levels of approved use, and non-adherence to syllabus or program statements can be treated as a potential academic integrity violation
In some courses, your instructors might encourage you to use these tools for specific purposes and assignments to help you develop awareness, skills, and knowledge of how they work. Other instructors might make it clear in the syllabus or in class that the use of generative AI is inappropriate for the assignments in that class. If you’re unsure about using generative AI for a particular assignment, it’s a good idea to talk to your instructor.
Graduate students must read and understand these program statements, and appreciate that individual faculty may have different levels of approved LLM use in their classes, ranging from no use of LLMs allowed, to students being encouraged to use LLMs freely. Non-adherence to syllabus and program-level statements on LLM use can be considered grounds for a report of a potential academic integrity violation.
- The university treats use of unauthorized aids, including electronic devices and online resources, on exams as academic dishonesty
- Departmental policies add stricter AI-specific rules in some units: Psychology says generative AI may not be used for exams unless expressly permitted by the instructor, and the CS department says students may still need to explain their entire solutions in a controlled setting
Cheating on exams or assignments by the use of books, electronic devices, online resources, notes, or other aids when these are not permitted, or by copying from another student.
Electronic communication devices, including cellular phones, speakers, calculators, electronic translators, smart watches, and headphones must be secured in a closed container (and not, for example, worn on a belt or around the neck) and must be turned off (and not, for example, simply set on vibration mode) during any examination.
● Generative AI or large language models (LLMs, such as ChatGPT) may not be used for exams, to satisfy academic milestones, or to produce assignments in psychology classes unless expressly permitted by the instructor.
Finally, whether AI is used or not, students should be able to stand in front of a class, instructors, and/or TAs, and explain their entire solutions and work in a controlled setting (e.g., like a closed-book exam).
- The Writing and Rhetoric graduate certificate also permits students to experiment with AI to assist writing and learning processes
- Psychology cautions that AI may be useful for exploring unfamiliar topics and reading lists, but warns students to be careful and remain responsible for output
- Graduate-level university guidance permits limited AI use for learning support tasks such as brainstorming, reading lists, and, in the CS department, exam preparation and finding initial information, with attribution
Guidelines to Programs about Potentially Appropriate Uses of LLMs (all with discipline specific appropriate attribution)
* Initial brainstorming
* Suggest improvements/edits to user-generated content
* Generation of reading lists
* Image generation
* Computer code testing
AI may be used for brainstorming ideas, for exam preparation, for finding initial information, for generating initial lists of ideas and/or related papers to investigate, and for some image generation.
The Department of Writing and Rhetoric permits students to experiment with AI to assist writing and learning processes.
LLMs might prove useful for exploring an unfamiliar topic, assembling reading lists, learning about a new format or genre such as a cover letter, drafting emails, creating stimuli for psychology experiments, proofreading, or writing code.
- The CS department prohibits AI as a wholesale substitute for coding, while Psychology notes AI may be useful for writing code
- In CS courses, actual permission still depends on instructor approval, with additional written approval required for undergraduate classes
- University graduate guidance allows computer code testing as a potentially appropriate use with discipline-specific attribution, but says complete problem solutions are inappropriate
Guidelines to Programs about Potentially Appropriate Uses of LLMs (all with discipline specific appropriate attribution)
* Initial brainstorming
* Suggest improvements/edits to user-generated content
* Generation of reading lists
* Image generation
* Computer code testing
Guidelines to Programs about Inappropriate Uses of LLMs
* Complete text generation, including documents to satisfy major academic milestones
* Complete problem solution
AI may not be used as a wholesale substitute for the work expected of students: solution generation, coding, or scholarly writing (e.g., writing major portions of essays and papers).
Undergraduate: AI may be used in undergraduate classes if and only if the instructor permits it in writing, and with prior written approval of the Undergraduate Program Director (UPD).
Graduate: AI may be used in graduate classes if and only if the instructor permits it in writing.
LLMs might prove useful for exploring an unfamiliar topic, assembling reading lists, learning about a new format or genre such as a cover letter, drafting emails, creating stimuli for psychology experiments, proofreading, or writing code.
Research
- Graduate School policy requires theses and dissertations to be written by the author; LLM-generated content is only considered when integral to the study and cannot form the majority of the document
- Departmental policies also require citation or acknowledgment for academic or research use, and the Psychology policy allows readability improvement with acknowledgment while keeping authors fully responsible for submitted or published material
All theses and dissertations must be written by the author. Exceptions to allow LLM generated content will only be considered if the LLM generated content is integral to the purpose of the study. However this type of content cannot make up the majority of the document attributed to the author. Theses and dissertations cannot list multiple authors, and this includes LLMs.
● Author(s) must take all responsibility for material they submit for publication. Generative AI cannot take responsibility for any material it generates.
● Academic or research use of generative AI technology must be cited. See suggestions below.
● Generative AI can be used to improve readability of text in some cases (although in doing so, it may lose information). This use must be cited in an acknowledgment.
- Psychology notes AI may be useful for creating stimuli for psychology experiments
- The Graduate School states that LLMs can be integral to academic research and scholarly activities and can facilitate data processing
- The Ecology & Evolution policy and Graduate School guidance emphasize that community members are responsible for ethical scholarship and must evaluate and correct any shortcomings in AI output, as those shortcomings become the user's own scholarly failures if uncorrected
We believe that LLMs can offer insight and inspiration into graduate level work, be integral to academic research and scholarly activities, facilitate data processing, and allow for new discoveries.
LLMs might prove useful for exploring an unfamiliar topic, assembling reading lists, learning about a new format or genre such as a cover letter, drafting emails, creating stimuli for psychology experiments, proofreading, or writing code.
Community members are responsible for ethical scholarship, and AI output does not necessarily meet that standard. Any shortcoming of AI output becomes shortcomings of the user's scholarly work if not corrected. The user is always responsible for evaluating AI output and identifying any errors or misrepresentations.
- No source specifically addresses grant proposals or IRB applications
- Departmental research-oriented policies require ethical responsibility, disclosure, and citation when AI is used in research contexts
- Ecology & Evolution requires clear description, disclosure, labeling of AI output, and documentation before publication; Psychology requires citation for academic or research use and states authors remain responsible for submitted and published work
The broad availability of generative AI tools is already driving changes in how both research and education are done. However, fundamental values of both education and research (such as promoting individual growth and mastery, honesty and responsibility, and reproducibility) will not change.
The graduate program in Ecology and Evolution is charged with overseeing the intellectual development of its students. To accomplish that goal, graduate program faculty must be able to identify and evaluate the students’ own intellectual work. Thus, while the program believes that the use of large language model generative AI tools may be acceptable in certain cases, the program also requires clear description of AI use, as with any tool. When their use is permitted, generative AIs are to be treated as tools, just like any tool used to do science. That means the use of an AI tool must always be disclosed, and any presentation of output from an AI tool be clearly labeled as such, to distinguish it from the human contribution.
Also, students should be cognizant of what types of restrictions or disclosures are used by journals, so they can adequately document use of large language models and generative AI to disciplinary standards prior to publication.
● Academic or research use of generative AI technology must be cited. See suggestions below.
● Author(s) must take all responsibility for material they submit for publication. Generative AI cannot take responsibility for any material it generates.
Academic Integrity
- Disclosure and attribution requirements are explicitly stated in several sources
- Graduate School guidance forbids unattributed use and calls for discipline-specific attribution; Ecology & Evolution requires disclosure of tool, dates, and prompts; the CS department requires attribution; CELT recommends citation and acknowledgment of AI-generated or AI-assisted content; and Psychology provides attribution language for academic or research use
However, LLMs should not be used in an unattributed manner, and/or as a substitute for your own scholarly writing.
Guidelines to Programs about Potentially Appropriate Uses of LLMs (all with discipline specific appropriate attribution)
The use of an AI tool must always be disclosed, and any presentation of output from an AI tool be clearly labeled as such, to distinguish it from the human contribution. If students use AI in any product used for assessment of their work in Ecology & Evolution, they are required to explain what permitted AI content-generation tool was used, the dates it was accessed, and the prompts (or types of prompts) used to generate the content, according to an appropriate style guide
AI use must be attributed (e.g., you must cite your sources and how they were used).
Citing AI Content: Any AI-generated content that is quoted, paraphrased, or otherwise incorporated into a student’s work should be cited including AI-generated images, media, and other visual content.
Acknowledgment of AI Assistance: If students are using AI tools to assist them with different learning activities such as research and composition, it is important that they know where and how to document this usage.
● Academic or research use of generative AI technology must be cited. See suggestions below.
Suggestions for how to attribute the use of AI (in lieu of guidance from course instructors or journals):
I acknowledge the use of <insert tool name(s), date/version, url>
- Graduate guidance encourages programs to include statements on AI detection software and notes such tools have limitations
- CELT provides a faculty process using Turnitin AI detection and emphasizes review for false positives, comparison to course policies, student discussion, and documentation
- The university's academic integrity process applies to suspected AI misuse: representing AI-generated work as one's own is listed as academic dishonesty, instructors must report suspicions to the Academic Judiciary Office, and penalties can include an F for the course or more severe sanctions
Representing work generated by artificial intelligence as one's own work.
Course instructors who suspect violations of academic integrity must report their suspicions to the Academic Judiciary Office; they may not establish a penalty independently.
An F for the course is considered to be an appropriate penalty grade for an academic integrity violation that occurs within the context of a course or courses. A more severe penalty may be recommended under certain circumstances.
We also encourage Graduate Programs to include statements regarding the use of Generative AI detection software, which can provide guidance on the likelihood of generative AI content, but have the same limitations as current LLMs and currently lag behind the advances of the LLMs generation tools themselves.
Examine the flagged paper within the Turnitin AI detection tool.
If the flagged content seems minor or likely a false positive, you may decide no further investigation is needed.
Compare the flagged sections with:
* Course assignment guidelines and requirements.
* Course AI policy in your syllabus.
* Known patterns or prior work submitted by the student (E.g., writing style or complexity).
Institutional & Administrative
- Additional faculty teaching resources are available through CELT's AI teaching resources page
- CELT provides guidance on evaluating AI-flagged work fairly and describes AI uses for course design such as developing learning objectives and multiple-choice questions
- Faculty-facing guidance asks instructors to define course AI policies, articulate expectations in syllabi and assignments, and address confidentiality, student responsibilities, and citation practices
As AI and machine-learning tools such as Microsoft Copilot, Google Gemini, and more continue to evolve and expand their capabilities, faculty are asked to consider the role these tools will play in their courses and how best to manage their use in ways that support teaching and learning.
* For what purpose will students be permitted to use these tools? Can they use them for brainstorming? Proofreading? Composing text?
* What will the expectations be for students who use AI?
* What do I consider to be academic dishonesty within my course with respect to generative AI usage?
* How can I ensure that students are informed of and respect any confidentiality and/or private policies?
* How will I ensure that students understand their responsibilities when working with AI-generated content?
* How will students be able to disclose and/or cite their usage of generative AI?
As the use of artificial intelligence tools in academic writing continues to grow, faculty should take a proactive and fair approach when evaluating AI-flagged student work.
Develop Learning Objectives: AI tools can help in the development of new learning objectives or to rework existing learning objectives.
Develop Multiple Choice Questions: AI tools can help in the development of multiple choice questions or to provide feedback on existing multiple choice questions.
Our Google Apps for Education license provides access to all to chat with Gemini at no additional cost. Faculty/Staff seeking the ability to use within Gmail, Docs, Vids and other Google Apps, your department will need to purchase a Google AI Pro license.
- The DoIT AI tools page identifies institutionally offered AI tools and states that listed tools may not be used with HIPAA data
- The university's data policy also limits access to university data to legitimate educational or institutional purposes and prohibits improper disclosure or re-disclosure
- Stony Brook has data protection rules that apply across the university, including the academic and research enterprise, and require approved minimum-security standards based on data classification
This policy applies to all members of the university community, including the academic and research enterprise, the health system, Long Island State Veterans Home, other units as may come under management of the University, and third parties that handle University Data.
Stony Brook classifies physical and electronic data into three risk-based categories for the purpose of determining access, permissions, and security precautions.
All University Data fall into one of the three categories. Based on the data classification, individuals who use University Data are required to implement approved minimum-security standards, where available, for protecting the data.
Access to University Data requires a legitimate educational or institutional purpose.
University data may not be improperly disclosed or re-disclosed.
A current directory of AI tools offered by DoIT for smarter meetings, ensuring academic integrity, organizing your thoughts, or simplifying your workflow
The following tools may not be used with HIPAA data.
- The group has executive sponsors from university leadership and members from multiple units
- The university states that it does not yet have a single AI policy and is reviewing existing policies
- Governance is being coordinated by the AI Coordination and Communication Work Group, a standing advisory group that coordinates AI-related initiatives, approaches, practices, and policies and communicates them to campus stakeholders
No. The University is reviewing existing policies to ensure they appropriately address the power and implications of generative AI. The AI Coordination and Communication work group is assisting policy owners to conduct this review.
The AI Coordination and Communication Work Group is a standing advisory group charged to coordinate information about initiatives, approaches, practices, and policies about AI and to communicate this information clearly to campus stakeholders. As AI matures, the role of the group may evolve at the discretion of the Executive Sponsors.
## Executive Sponsors
* Chief Deputy to the President and Senior Vice President
* Senior Vice President for Finance and Administration
* Provost and Executive Vice President for Academic Affairs
* Executive Vice President for Stony Brook Medicine
* Vice President for Information Technology and Chief Information Officer
* Vice President for Educational and Institutional Effectiveness
DocuMark: Responsible AI Use for Academic Integrity
Knowing your institution's AI policy is step one. DocuMark helps enforce it fairly by empowering universities to manage AI-generated content, prevent cheating, and support student writing through responsible AI use.
Common Questions About Stony Brook University--SUNY's AI Policies
Stony Brook University--SUNY has defined AI policies in 12 of 12 categories, with an overall coverage score of 100%.
Disclosure and attribution requirements are explicitly stated in several sources. Graduate School guidance forbids unattributed use and calls for discipline-specific attribution; Ecology & Evolution requires disclosure of tool, dates, and prompts; the CS department requires attribution; CELT recommends citation and acknowledgment of AI-generated or AI-assisted content; and Psychology provides attribution language for academic or research use.
The university's academic integrity process applies to suspected AI misuse: representing AI-generated work as one's own is listed as academic dishonesty, instructors must report suspicions to the Academic Judiciary Office, and penalties can include an F for the course or more severe sanctions. Graduate guidance encourages programs to include statements on AI detection software and notes such tools have limitations. CELT provides a faculty process using Turnitin AI detection and emphasizes review for false positives, comparison to course policies, student discussion, and documentation.
Stony Brook has data protection rules that apply across the university, including the academic and research enterprise, and require approved minimum-security standards based on data classification. The DoIT AI tools page identifies institutionally offered AI tools and states that listed tools may not be used with HIPAA data. The university's data policy also limits access to university data to legitimate educational or institutional purposes and prohibits improper disclosure or re-disclosure.