����ֱ��

Background

Since 2023, Queen’s University has maintained an institutional position on the use of generative artificial intelligence in teaching and learning. That position established a framework in which instructors determine whether and how AI tools may be used in their courses, and in which unauthorized use of AI constitutes a departure from academic integrity. Separate guidance issued in 2024 addressed the use of generative AI by instructors and teaching assistants for the purposes of student assessment. These positions remain in effect.

This statement addresses a new development in the AI landscape: the emergence of agentic AI.

What is agentic AI?

Agentic AI tools can act autonomously on behalf of a user: navigating websites, logging into systems, completing multi-step tasks, and making decisions without continuous human direction. Some tools have been designed to interact with learning management systems and submit work on a student’s behalf. Some run locally on a user’s device, making them more difficult to detect or assess through standard institutional security processes.

How agentic AI differs from generative AI, and why this matters

Assessment is not only a mechanism for assigning grades; it is how students consolidate learning, receive feedback, and demonstrate developing competency. When an autonomous agent completes that work, the intended opportunity for learning is all but eliminated. The student is not engaged more efficiently; they are not at all engaged with the assessment’s intended learning outcomes.

Generative AI tools such as ChatGPT produce content in response to a user’s prompt, but the user retains control. When a student uses a generative AI tool to produce work, at least some decisions remain theirs: what to prompt, how to revise and what to submit.

For these reasons, using agentic AI to complete coursework is functionally indistinguishable from contract cheating. Queen’s defines contract cheating in the Academic Integrity Procedures as outsourcing academic work to a third party and submitting it as one’s own. The fact that the third party is a software agent does not change the nature of the act; the departure lies in the delegation.

Treating agentic AI misuse as a form of contract cheating is accurate as a matter of policy, but the analogy has real limits in practice. The academic integrity policy and procedures were designed for departures that leave detectable evidence. Agentic AI misuse typically does not. Therefore, for most instructors, in most courses, unauthorized agentic AI use will not be identifiable after the fact.

The guidance that follows outlines what instructors and the institution can do. Detection after the fact is not a reliable response; the more durable answer lies in how assessments are designed and what evidence they are built to produce.

For instructors

The following extends the university’s existing position on AI in teaching and learning. It does not replace the guidance issued in 2023 and 2024.

Using agentic AI as a teaching tool

Where the intentional, supervised use of agentic tools is itself the object of learning, such use likely has pedagogical value. For example, an assessment that asks students to configure, deploy, and critically evaluate an agentic tool engages students directly with the technology rather than delegating work to it. Instructors considering such approaches should ensure their course outline explicitly defines the scope and purpose of that use and should not require students to use any Queen’s credentials or personal information (e.g., personal email address, phone number) to access any agentic tool.

Suggested syllabus language for both permitted and non-permitted agentic AI use is available on the Centre for Teaching and Learning website.

Assessment design

Assessments completed online without supervision and without interactive or iterative components are particularly susceptible to completion by agentic AI.

Where such assessments contribute to a final grade, instructors are encouraged to consider whether these assessments remain appropriate mechanisms for evaluating individual student learning. Using unsupervised assessments for formative purposes, where the goal is feedback rather than evaluation, remains appropriate.

Addressing unauthorized use

The use of agentic AI tools to complete academic work on behalf of a student constitutes a departure from academic integrity, regardless of how an instructor has framed the permitted use of generative AI in a course. A syllabus statement permitting the use of generative AI tools does not extend to the offloading of coursework to an autonomous agent. Doing so falls within existing categories of departure, including unauthorized content generation and contract cheating, as defined in the Academic Integrity Procedures.

Unlike some other departures from academic integrity, unauthorized agentic AI use is very likely undetectable through the review of submitted work alone. The most reliable instructor-level response is in the course design process: designing assessments where the student’s process is visible or verifiable, not just their product.

Where an instructor has reason to believe a student has used agentic AI to complete work in an unauthorized way, the process of investigation is the same for any other suspected departure from academic integrity. If you have questions about how you might proceed with such an investigation, please reach out to your Academic Integrity Administrator.

For Faculties and Schools

Assurance of learning at the program level

Agentic AI reinforces the importance of assurance of learning: ensuring that assessments provide credible evidence that students have developed the knowledge, skills, or values a program is designed to build. The degree of assurance required varies, where some assessments are primarily formative, while those that mark achievement of a program-level learning outcome warrant stronger evidence of individual student learning. Determining where that assurance is needed is a programmatic discussion.

Faculties and Schools are now even more strongly encouraged to consider how program-level assessment planning can support the integrity of their programs; the Centre for Teaching and Learning is available to support this work.

University-level response

IT Services is actively assessing the risks that agentic AI tools pose to institutional systems, including the learning management system, and is engaging with vendors to address those risks at the platform level.  If you are considering using a new AI tool, you are encouraged to see if has been reviewed by ITS.

The VP TL portfolio will continue to work with IT Services to ensure that pedagogical and security responses are coordinated.

These existing policies or procedures are relevant to the use of both generative and agentic AI tools in teaching and learning:

Academic Integrity Procedures: The use of agentic AI to complete coursework on a student’s behalf falls within existing definitions of departure from academic integrity. No changes to the Academic Integrity Procedures are required to address this; the existing framework applies. Students found to have used agentic AI in this way are subject to the same processes and consequences as other academic integrity matters.

Course Outline Policy: The recently approved course outline policy requires, effective September 2026, that instructors indicate in the course outline whether and how AI tools may be used in the course, regardless of the type of AI involved. See the Course Outlines Standards and Guidelines for more information.

Copyright Compliance and Administration Policy: The use of generative or agentic AI tools does not override existing copyright obligations. Faculty, staff, and students remain responsible for ensuring that materials uploaded to, generated by, or reused through AI tools comply with applicable copyright requirements and licensing terms.

Access to Information and Protection of Privacy Policy: The use of generative or agentic AI tools may involve the collection, use, disclosure, or handling of personal information or other university records. Existing privacy obligations continue to apply. Faculty and staff are responsible for ensuring that personal and confidential information in the university’s custody or control is handled and protected in accordance with the policy, FIPPA, and other applicable requirements.

Responsible Use of Digital Resources Policy: Some agentic AI tools may interact with university systems, accounts, files, data, or digital services on a user’s behalf. Existing expectations regarding the responsible, ethical, lawful, and secure use of university digital resources continue to apply. Users remain responsible for protecting university credentials and data in their care and control.

Academic Accommodations for Students with Disabilities Policy and Procedure: Decisions about the use of AI tools in teaching and learning should be made in a manner consistent with the university’s academic accommodation framework. Some AI-enabled tools may be institutionally approved for accommodation purposes. Where this is the case, their use is governed through existing accommodation processes and institutional approvals, while maintaining essential academic requirements and standards. Course-level decisions about whether AI tools are restricted, required, or permitted should therefore take account of accommodation obligations and should not be interpreted to displace approved accommodations.

This statement will be reviewed and updated as the technology, the risk environment, and institutional practice evolve.

Gavan Watson,

Vice-Provost, Teaching and Learning

As we enter the 2024/2025 academic year, please note that that the guidance on generative artificial intelligence (GenAI) communicated in September 2023 remains unchanged. Key points include:

• No broad ban on GenAI tools: Queen’s has not banned generative AI tools.
• Provide students guidance: Instructors should specify the parameters for AI tool use in their courses and advise on terms of use via a syllabus statement.
• Academic integrity: Unauthorized use of generative AI is considered a departure from academic integrity.
• AI-detection tools: These tools should not be used due to privacy and reliability concerns.

Why Generative AI Tools Are Not Banned

Due to their widespread presence and the challenges in detecting their use, Generative AI tools are not banned at Queen’s. The university also acknowledges the persuasive arguments regarding the potential benefits these technologies can offer.

Provide students guidance on use

Instructors are expected to clarify if and how generative AI can be used in their courses, and it’s recommended to discuss their stance with students.

Include a syllabus statement

Instructors should include clear guidelines or a statement in the syllabus to advise students on the use of generative AI tools. The below statements were developed by the Faculty of Engineering and Applied Science and shared with permission. 

• Permitted with citation
  • Students must submit their own work and cite the work that is not theirs. Generative AI writing tools such as ChatGPT are welcome in this class, provided you cite the material that they generate. Any other use constitutes a departure from academic integrity.
• Permitted in specific assignments, with citation
  • Students must submit their own work and cite the work that is not theirs. Generative AI writing tools such as ChatGPT are only permissible when explicitly noted in the assignment instructions. In these cases, be sure to cite the material that they generate. Any other use constitutes a departure from academic integrity.
• Not permitted
  • Using generative AI writing tools such as ChatGPT in your submitted work is not permitted in this class. This type of use constitutes a departure from academic integrity. Original work, completed wholly by you, is expected to be submitted in this course.

Unauthorized use of GenAI is a departure from academic integrity

At the October 5, 2023 Senate meeting, Senate approved proposed revisions to the Academic Integrity Procedures. These revisions introduced a new type of departure from academic integrity called Unauthorized Content Generation and modifying existing categories like plagiarism to include the misuse of technological assistance.

In April, 2024 the Office of the Provost and Vice-Principal (Academic) developed a resource titled “5 investigations steps for instructors for suspected unauthorized content generation,” which was shared with Associate Deans in Faculties and Schools for broad distribution.

Avoid AI detection tools

The use of third-party AI-detection tools is strongly discouraged due to their unreliability and the potential breach of student privacy and intellectual property rights.

Ongoing monitoring and support

The Vice-Provost, Teaching and Learning portfolio will continue to monitor the influence of AI on education, providing support and guidance to the institution, academic units, and instructors.

Further support

For further guidance around teaching and curriculum design, please review the resources created by Centre for Teaching and Learning.

July 2024

As educators we share ethical, legal, and professional obligations to support the success of all learners at Queen’s University. Through understanding the foundation of our institutional practices, Queen’s educators and staff are in a better position to help enhance the learning environment for all. This statement aims to set the foundation for academic accommodation practices for use by instructors and staff involved in instruction at all levels at Queen’s University.

It summarizes:

1. the provincial legal context for academic accommodation;
2. how academic accommodations are made;
3. the role (and limits) of “universal” frameworks in supporting student accommodation; and
4. a listing of supporting services and resources on accommodation.

1. The Legal Context for Accommodation

The Accessibility for Ontarians with Disabilities Act (“AODA”) and the Ontario Human Rights Code (“the Code”) are laws that work together with the goal of making the province more inclusive by reducing and removing barriers. It is under the AODA and the Code that the institution has a legal obligation to provide academic accommodations to students.

In (“the report”), the Ontario Human Rights Commission (“OHRC”) establishes a foundation of principles for student accommodation in post-secondary education. In the report, guidelines detailed in the following sections are especially relevant for our work as an institution:

As detailed by the OHRC, accommodations should not lower academic standards or interfere with the achievement of different outcomes. Through the development of flexible engagement in the academic material (including tests, quizzes, and exams), appropriate accommodation removes or reduces barriers experienced by a student with a disability, allowing them the equal opportunity to demonstrate individual mastery of the academic material.

Post-Secondary Institutions

The duty to accommodate rests on the whole educational institution with responsibilities shared by instructors, the institutional office for students with disabilities, and institutional policy makers. Students with disabilities engage with the university in the accommodation process.

Post-secondary institutions are responsible for ensuring that their facilities and services are accessible; that the environment is welcoming and non-discriminatory; that appropriate, effective, and dignified accommodation processes are in place; and students who experience functional impacts in a classroom because of their disabilities are provided an appropriate accommodation to the point of undue hardship[1] for the institution.

As detailed in the report, all members of the post-secondary institution have a role to play.

For example,

• Faculty and staff have a duty to educate themselves about disability-related issues and the processes and guidelines for seeking accommodation, to interact with students in a non-discriminatory manner, to engage in the accommodation process, and to provide appropriate accommodation to the point of undue hardship.
• Staff and faculty responsible for designing or developing new or revised facilities, services, policies, processes, courses, or curricula[2] must ensure that these are designed inclusively, with flexibility to allow students with disabilities access to the course with minimal impact to curriculum or student.
• Clear and reasonable processes and guidelines for seeking accommodation should be in place at all post-secondary institutions, and these should be clearly communicated to all students.
• Based on the three principles of the duty to accommodate as outlined by the OHRC in the report the process of accommodation, as well as the outcome, should be respectful of the dignity of the persons seeking accommodation, engage with the person seeking accommodation as a unique individual with unique accommodation requirements, and should consider the importance of integration and full participation.
• Any university plans and policies related to accessibility should recognize that persons with disabilities are important stakeholders in the process.

Post-secondary institutions have a responsibility, short of undue hardship, to cover the cost of the required accommodation, unless there is sufficient (and non-discriminatory) outside funding available.

In 2016, the OHRC outlined further specific measures for universities to implement with the goal of reducing systemic barriers to post-secondary education, while ensuring that institutions receive appropriate documentation to determine academic accommodations. These measures further guide our institutional processes and detail our obligations, specifically through the Academic Accommodations for Students with Disabilities Policy. The measures detailed by the OHRC, via a 2017 report () are as follows:

• When seeking an accommodation for functional impacts related to a mental health disability, students must provide information about their disability-related needs, but it is not mandatory for them to disclose a medical diagnosis.
• Students are not required to reveal private medical information to faculty or instructors for the purpose of having their approved accommodations implemented in a class.
• Interim, retroactive, permanent, or temporary accommodations are valid for both temporary and permanent disabilities.

• Rather than directly from instructors, students should seek accommodation through a centralized process.
• Guidelines, forms and procedures are to be clearly communicated to students, faculty and staff.

In support of these measures, Queen’s students, faculty and designated staff can access approved student accommodation information via Ventus, the University’s accommodation management portal. While students’ accommodations are approved though a central process, the implementation of the accommodations in individual courses requires the engagement of instructors, QSAS staff and students.

2. Meeting our Duty to Accommodate

Accommodations are intended to provide students with disabilities[3] equal learning opportunities and equal access to the learning environment. Accommodations are not designed to give a student an advantage over other students, to alter essential requirements of a course, nor to weaken academic rigor. Rather, the duty to accommodate within the context of a Queen’s University classroom or course means that the institution, through shared responsibility and through a process outlined in part below, adjusts the learning experience so that students with disabilities can fully participate, unless that accommodation causes undue hardship.

With education, training and experience related to disabilities and the barriers students face during the learning process, staff within the Queen’s Student Accessibility Services (“QSAS”) office seek to remove disability related academic barriers while working with instructors to uphold essential program requirements and are committed to supporting students as they pursue their academic goals.

The QSAS process comprises a review and assessment of legislative requirements and institutional policy, along with understanding of the first-person experience of disability impacts from the individual student, and rigorous documentation review. The process for implementing accommodations is actioned in the following way:

• Student registers with the Ventus portal using their NetID and password where they are prompted to fill out an intake form.

• The student reviews the QSAS Documentation Criteria and uploads the appropriate documents to Ventus.

• Intake Coordinators provide the first level of understanding the students’ disability related needs by reviewing the first-person experience outlined in the intake form and reviewing the provided documentation. A determination is made whether the student is eligible for permanent, temporary, or interim accommodations.

• Accessibility Advisors further review the students’ first-person experience and the provided documentation to identify specific functional impacts and barriers that the student may face.

• Accessibility Advisors and students work collaboratively to understand the unique student barrier(s) and implement appropriate accommodations that remove the disability-related barrier while also upholding the essential requirements of the course.

• Accommodations are provided to the student through Ventus.

3. Academic Accommodation and Universal Design for Learning

With the goal of improving student outcomes, (“UDL”) is a structured framework that calls for instructors to provide multiple ways of presenting classroom material, multiple ways of engaging the student in the learning, and multiple ways for the student to demonstrate learning. Universal Instructional Design (“UID”) is a framework intended to support the implementation of UDL by developing flexibility within the curriculum of a course.

While UDL and UID frameworks encourage educators to design learning experiences to help reduce barriers to student learning from the design or planning stage of a course, it is important to keep in mind the following from the perspective of student accommodation:

• UDL and UID approaches will not offer a substitute for all learner needs, as while these approaches can offer more inclusive classrooms, not all UDL/UID strategies can meet accommodation needs of all students.

4. Additional Resources on Accommodation

����ֱ�� Student Accessibility Services (QSAS)

Queen’s Accessibility Hub: Students

• Overview for Students
• Documentation and Accommodations
• Supports and Services for Students with Disabilities
• Human Rights and Equity Office
• Medical professionals (from Student Wellness Services or from the community or students’ regular family physician office in their hometown)
• Academic advisor / faculty advisor
• (SASS)

Queen’s Accessibility Hub: Instructors

• Teaching Resources
• Accommodating Students with Disabilities
• Education and Awareness – Training for Faculty and Staff
• Accessible Instruction for Educators

Human Rights and Equity Office

Ontario Human Rights Commission:

Last updated: September 19, 2023

In anticipation of the 2023-24 academic year, please find an overview of the guidance and updates communicated by the Vice-Provost, Teaching and Learning since February 2023, specifically as related to addressing the impact of generative artificial intelligence (tools such as ChatGPT) within the classroom environment.

Queen’s has not banned the use of generative AI tools. This technology has already entered the educational system, is very difficult to detect, and is so widespread that it would be difficult, if not impossible, to prevent its use. In addition, many persuasive arguments have been made about the potential positive uses of generative AI.

(from the General Statement from the Vice-Provost (Teaching and Learning) on ChatGPT and Generative AI, February 2023)

However, inappropriate use of generative AI constitutes a departure from academic integrity. Among the core values of academic integrity are honesty and fairness that establish a framework for teaching and learning for both undergraduate and graduate students at Queen’s.

(from the General Statement from the Vice-Provost (Teaching and Learning) on ChatGPT and Generative AI, February 2023)

To update academic integrity policy in light of generative AI, the Academic Integrity Subcommittee of the Senate Committee for Academic Development and Procedures has drafted updates to the types of departures from academic integrity in the context of generative artificial intelligence. These are expected to go forward for approval at the first Senate meeting of the academic year (October 2023).

Instructors should indicate if this technology can be used in a course and if so, what the parameters of its use will be. Generative AI has distinct impacts on course assessment practices. Instructors have the most immediate sense of the potential impacts on their courses, and they are best placed to make key decisions about whether generative AI fits with their learning outcomes. It is essential, then, that instructors speak to their classes about their perspectives on generative AI and detail the conditions for its use in their courses.

(from the General Statement from the Vice-Provost (Teaching and Learning) on ChatGPT and Generative AI, February 2023)

Instructors should include a syllabus statement to advise students on the use of generative AI tools. The first three statements were developed by the Faculty of Engineering and Applied Science and shared with permission. 

• Permitted with citation:
  • Students must submit their own work and cite the work that is not theirs. Generative AI writing tools such as ChatGPT are welcome in this class, provided you cite the material that they generate. Any other use constitutes a departure from academic integrity.
• Permitted in specific assignments, with citation:
  • Students must submit their own work and cite the work that is not theirs. Generative AI writing tools such as ChatGPT are only permissible when explicitly noted in the assignment instructions. In these cases, be sure to cite the material that they generate. Any other use constitutes a departure from academic integrity.
• Not permitted:
  • Using generative AI writing tools such as ChatGPT in your submitted work is not permitted in this class. This type of use constitutes a departure from academic integrity. 
  • Original work, completed wholly by you, is expected to be submitted in this course. The use of an artificial intelligence tool like ChatGPT is not permitted.

(from Update to Faculty Deans and Associate Deans (Teaching and Learning) from Vice-Provost (Teaching and Learning) on Turnitin's new AI Detection Tool, April 2023)

Third-party tools designed to detect AI-generative text should not be used to check student work. No current evidence exists that demonstrates that any third-party AI-detection tool reliably works to detect text generated by artificial intelligence. Submitting students’ work to AI detection software without student permission is a breach of the student’s privacy and intellectual property, and as such third-party AI-detection tools should not be used.

(from Update to Faculty Deans and Associate Deans (Teaching and Learning) from Vice-Provost (Teaching and Learning) on Turnitin's new AI Detection Tool, April 2023)

As we enter a new academic year, the influence of artificial intelligence tools on teaching and learning will continue to be an area of focus for the Vice Provost, Teaching and Learning portfolio. We will continue to engage in consultation and research with the aim to advise and support the institution, academic units and instructors on the implications and opportunities presented when using this technology in our classrooms.

For further review, the Centre for Teaching and Learning has developed further resources introducing large language AI models and their impact on teaching.

August 2023.

General Statement from the Vice-Provost (Teaching and Learning) on ChatGPT and Generative AI - February 22, 2023

Discussions of the impact of ChatGPT and generative AI technologies range from those heralding the opening of an entirely new era of educational innovation to those condemning the emerging technology as something which will bring about the end of the essay, undermine the development of creative thinking and originality, and lead to an epidemic of plagiarism. What is clear from reading even a small sample of these essays, blogs, tweets, and other commentaries is that we must confront, understand and attempt to manage the use of this technology in higher education.

While I am beginning formal discussions across the Faculties and Schools, it is apparent that many members of the ����ֱ�� community are looking for some general guidance on our approach to the management of these generative AI technologies. In discussions with Deans, Associate Deans, instructors and staff, I would like to suggest the following principles to guide our initial approach to generative AI:

• Queen’s will not ban the use of generative AI technologies.
  • This technology has already entered the educational system, is very difficult to detect, and is so widespread that it would be difficult, if not impossible, to prevent its use. In addition, many persuasive arguments have been made about the potential positive uses of generative AI
• Instructors should indicate if this technology can be used in a course and if so, what the parameters of its use will be.
  • It is clear that generative AI has distinct impacts on course assessment practices. Instructors have the most immediate sense of the potential impacts on their courses, and they are best placed to make key decisions about whether or not generative AI fits with their learning outcomes. It is essential, then, that instructors speak to their classes about their perspectives on generative AI and the conditions for its use in their courses.
• Inappropriate use of Generative AI would constitute a departure from academic integrity since it involves a misrepresentation of the student’s work and abilities.
  • Among the core values of academic integrity are honesty and fairness that establish a framework for teaching and learning for both undergraduate and graduate students at Queen’s. Honesty is manifest in "presenting one's own academic work" and "acknowledging dependence on the ideas or words" of any other source. Fairness involves a "full acknowledgement" of sources. The use of generative AI instruments without the consent of the instructor or proper acknowledgement of these sources compromises the foundation for a community of open exchange of ideas. In the terms of the departures from academic integrity, it also can potentially be considered a "use of unauthorized materials," depending on how the instructor frames the course requirements. (See the Academic Integrity Policy Statement and the Academic Integrity Procedures -- Requirements of Faculties and Schools).

More discussion will follow, and there will be further consultation across the university community. It is essential, however, that we all learn as much as possible about generative AI and consider the immediate implications for our courses. I know that the Deans and Associate Deans in each Faculty and School are developing approaches to the use of generative AI as it impacts their particular disciplines.

John Pierce, Vice-Provost (Teaching and Learning)