Guidelines for all uses of GenAI in Teaching & Learning 

These guidelines, developed with the Generative AI Teaching and Learning Advisory Subcommittee, provide guidance on responsible and ethical use of GenAI in teaching and learning at UBC.

1a. Developing literacy about GenAI

It is important for students, faculty, and staff to develop foundational literacy about GenAI tools and platforms, to support responsible, ethical, and effective use. This includes information about functionality, possible opportunities and benefits for teaching and learning, as well as potential challenges such as how they can be used to commit academic misconduct, how they can violate privacy, intellectual property and Indigenous data sovereignty rights, and risks of reproducing and perpetuating bias and inequities. Such learning takes time and effort, and will need to be gradually built up over time and updated as technology changes.  

UBC has resources available for learning about and experimenting with GenAI. These are updated with new information and events as they become available, and further resources will continue to be developed. 

1b. Academic integrity 

Permitted use 

  • Instructors, supervisors, and programs should clearly communicate to students the permitted and/or prohibited use of GenAI in their academic work, and the rationale for these decisions. Student use of GenAI outside of these stated rules may be considered academic misconduct. Please see the UBC academic integrity website FAQ on generative AI for details. 
    • Consider also that there are various kinds of GenAI, including functionality built into existing applications such as grammar checkers, word processing tools, and others. It is helpful for educators to be as specific as possible in what kinds of GenAI functionality is or is not permissible.  

Syllabus statements 

  • It is important for instructors to set clear expectations around the use of GenAI tools in the syllabus, as with any other tool or mode of working (group work, etc.), and reinforce this messaging throughout the term. Students may be navigating differing levels of GenAI permissions in multiple courses so communicating expectations in a clear and straightforward manner is important. 
    • The UBC Academic Integrity website has sample syllabus statements that can be used to communicate to students permitted and prohibited use of GenAI in courses. 

Citation and attribution 

  • If GenAI tools are allowed for student academic use, educators should make it clear to students when and how they should acknowledge its use, such as through citation, and students should follow those guidelines.   
    • Educators should consider the range of GenAI tools available, including that some are embedded in other tools such as grammar checkers, and decide whether and what kinds of uses of GenAI should be cited, and make that clear to students. 
    • If students are not sure whether and how to acknowledge use of GenAI for their academic work, they should reach out to their instructors or supervisors to clarify expectations. 
    • The UBC Library’s LibGuide on Generative AI provides information on how to cite the use of GenAI tools in multiple citation styles. 

Detection 

  • The use of applications to detect AI-generated content is strongly discouraged at this time, due to concerns about effectiveness, accuracy, bias, privacy, and intellectual property. 
  • Instructors should not submit student academic work or personal information to AI detectors that have not undergone a UBC Privacy Impact Assessment (PIA) and been approved for this use.  
    • If any detectors are approved for use at UBC in the future, it is recommended that faculty include information in their syllabi stating that such tools may be used in the course. 
    • As of June 2024, no AI detection tools have undergone a UBC PIA. The AI in Teaching and Learning website has a web page on current status of PIAs for generative AI tools at UBC. 
    • The announcement about UBC’s decision not to enable the AI detection feature in TurnItIn provides further information about concerns with AI detectors.  
  • Instead of AI detectors, some of the approaches to evaluating possible academic misconduct that educators have used before the broad availability of GenAI may also be effective when considering inappropriate use of GenAI, such as comparing work produced out of class to work produced in class, or having a conversation with students about their work.   

Academic Integrity processes 

  • If instructors or supervisors suspect that GenAI has been used inappropriately by students for their academic work, they should follow existing academic misconduct processes.  
  • Students who have questions or concerns about GenAI and academic misconduct, and who would like support and/or advocacy, can reach out to the following: 

Designing assessments to support academic integrity 

  • Educators can test the resiliency of their own assignments and assessments against the capabilities of GenAI tools if they choose, by submitting their assignment instructions to a GenAI platform to understand its capabilities and limitations. 
    • If educators decide to carry out such tests, they should only submit material that is their own, or for which they have received express permission to submit for this purpose. 
    • Before doing so, they should consider that this content could be used to train GenAI tools and could emerge publicly, depending on the tool used. 
  • The AI in Teaching and Learning website and the New frontiers: Teaching with AI blog provide information on resources, events, and other supports related to designing assignments and assessments that either use GenAI or that mitigate against its use for academic integrity purposes. 
  • Some educators may choose to have more in-class assessments to mitigate the use of GenAI. The CTLT at UBCV maintains a resource (jointly created with the UBCV Centre for Accessibility) on accommodations considerations for in-class assessments, which may be helpful for those considering this option. 

1c. Indigenous data sovereignty, protocols, and risks of harm 

There are diverse views of GenAI amongst Indigenous peoples and communities; the information and guidelines in this section should not be taken as representative of the views of all Indigenous people regarding GenAI. 

  • Indigenous Strategic Plan: Use of GenAI tools should uphold UBC’s commitments, goals, and actions in the current UBC Indigenous Strategic Plan; ensuring our work supports these is the responsibility of all at UBC. 
  • Indigenous data sovereignty & community protocols:  
    • Indigenous peoples have the right to steward their data according to their own community protocols. GenAI poses a risk to Indigenous data sovereignty by reproducing Indigenous knowledges, information, and data without the express permissions and protocols of Indigenous communities. 
    • There are several principles and practice guidelines related to Indigenous data sovereignty, including the First Nations OCAP principles (Ownership, Control, Access, and Possession), the National Inuit Strategy on Research the Principles of Ethical Métis Research, and the Indigenous rights in data from the Global Indigenous Data Alliance. Please note that these resources have been created by Indigenous organizations but should not be taken as representative of the governing protocols of specific Indigenous communities.  
    • GenAI tools that are trained on and that produce information that is owned by Indigenous peoples and communities violate Indigenous data sovereignty. In their very design and development, GenAI tools do not uphold the UN Declaration on the Rights of Indigenous Peoples (including Articles 3 and 4, on the rights to self-determination and self-government), which has been adopted into Candian federal law (2021) and BC provincial law (2019). In addition, UBC has committed to upholding UNDRIP through the Indigenous Strategic Plan (2020). 
    • Using GenAI to generate learning materials may violate Indigenous data sovereignty, such as by inputting Indigenous works into a GenAI tool without consultation or consideration of how the data will be stored, used, and shared. Using content generated by GenAI that includes or is based on Indigenous data may also violate Indigenous data sovereignty.  
  • Harmful outputs: GenAI tools can produce false, stereotyped, discriminatory and otherwise harmful information about Indigenous peoples and communities.  
  • Critical and responsible use: Below are considerations to help address the risks involved with GenAI tools and Indigenous knowledges, information, and other data. Given the concerns noted above these risks cannot be alleviated entirely. 
    • Learn about and follow principles related to Indigenous data sovereignty in any use of GenAI, such as those noted above. This includes:    
      • Do not input into GenAI tools any data, knowledge, or creative works by Indigenous peoples or communities without their express permission.  
      • Be aware that outputs from GenAI may share Indigenous knowledges outside the control or protocols of Indigenous communities. Instead of using and sharing information about Indigenous peoples and communities from GenAI tools, seek information from Indigenous sources. 
    • Take responsibility for your use and/or sharing of materials about Indigenous peoples generated by GenAI, and for the possible harms that may result. 
    • Do not rely on outputs from GenAI in efforts to include Indigenous content, knowledges, methodologies and approaches in teaching and learning.  
    • Do not use GenAI tools to learn about Indigenous histories and knowledges.  
      • There are many opportunities at UBC to learn, through resources, workshops, and events. Prioritize developing relationships with others as you learn and implement strategies for decolonizing teaching and learning, including participating in workshops and events led by Indigenous-focused units at UBC as well as public events led by Indigenous communities. 
         

1d. Ethical considerations

  • UBC plans and commitments: Use of GenAI tools should uphold UBC’s commitments, strategies, and actions related to equity, anti-racism, accessibility, sustainability, and wellbeing as laid out in multiple plans and frameworks, including the StEAR framework, the Climate Action Plan 2030, the Climate Emergency Task Force Report, the Wellbeing Strategic Framework, and the UBC Accessibility Plan (in progress).
  • Ethical benefits of GenAI: In some cases, GenAI can support inclusion, accessibility, and other ethical considerations. For example, GenAI tools can support learners with disabilities by being used to produce information in visual or audio formats in addition to text, to create captions for videos and alternative text for images, to help with focus and organization of writing or other tasks, and more.
  • Biased and harmful outputs, reproducing systemic inequities: GenAI tools are trained on large amounts of data that reflect social and cultural biases and dominant norms and values. These biases, and even discriminatory content that gets past guardrails built into many GenAI tools, can appear in outputs. Use of GenAI tools therefore can reproduce and further entrench systemic inequities.
  • Keep these considerations in mind when deciding whether or how to use GenAI tools in teaching and learning activities, and talk with students about these risks. 
  • Harm from false information: GenAI tools can produce confident-sounding but false information that in some cases can lead to harm to individuals or communities.  
    • Do not assume that what is produced by GenAI tools is true; double check the information by consulting other sources. 
    • UBC Librarians can support students with information literacy, including teaching how to critically evaluate false or misleading outputs from GenAI and working with instructors to provide course-specific training for students in information literacy and other topics. See the UBC Library’s resource on evaluating information from generative AI and the UBC Library Instruction webpages, UBCV and UBCO, for more information. 
  • Human oversight and responsibility: Those who use GenAI tools in teaching and learning should review outputs for possible harms to the best of their ability before sharing with others and take responsibility for redressing or mitigating harms that may result from the use of GenAI-generated content including potential violations of UBC’s commitment to a respectful environment. 
  • Inequitable access: It is important to recognize that UBC community members will have varying levels of access to GenAI tools, whether due to bandwidth, geographic location, costs for more powerful models, digital inaccessibility of some tools and platforms, or for other reasons.  
    • Keep this in mind when deciding whether and how to use GenAI in teaching and learning activities (also discussed further below under “Teaching with GenAI”). 
  • Sustainability: Training and using GenAI, depending on the models used, can lead to consumption of large amounts of energy and water. To support UBC’s commitments to sustainability and addressing the climate emergency, consider only using GenAI when doing so will have a significant benefit to teaching and learning activities, limiting use where it will be less impactful.  

1e. Privacy, security, confidentiality 

Many of the guidelines below are based on the Generative AI tools guidelines from the UBC Privacy Matters website. Note that privacy and confidentiality guidelines are based on legal and/or UBC policy requirements. 

  • Privacy Impact Assessments (PIA): In line with the British Columbia Freedom of Information and Protection of Privacy Act (FIPPA), instructors or teaching assistants cannot require that students use GenAI or any other technology tool that requires them to share personal information (such as name, student number, personal email address or phone number) unless that tool has undergone a UBC PIA review and been approved for use with personal information. 
  • This includes information such as names and personal email addresses of students, faculty, and staff; student numbers; and grades attached to identifiable students. It also includes anything that you would not, or do not have permission to make public, such as exam questions or other confidential data or materials. 
  • Privacy & confidentiality: In line with FIPPA and UBC Information Security Standards, do not share any personal, private, or confidential information when interacting with GenAI tools that have not undergone a PIA review and have not been approved for use with such information, as this data may be available to vendors, could be used for training models, and could end up in later outputs.
    • This includes information such as names and personal email addresses of students, faculty, and staff; student numbers; and grades attached to identifiable students. It also includes anything that you would not, or do not have permission to make public, such as exam questions or other confidential data or materials.
    • Most GenAI tools can only be used in teaching and learning at UBC with low risk information. As of May 2024, no GenAI tools have been approved for use at UBC with personal or other sensitive information.  
  • Terms of Service: It is recommended to review the Terms of Service and Privacy Policies of any GenAI tools you plan to use for teaching and/or learning, so you can decide whether you are comfortable with what data is collected and how it is used. Instructors should also discuss these with students if asking them to use GenAI tools.  
  • Use of data for training: Some GenAI tools offer the possibility of opting out of use of one’s input data for training models, which is recommended for privacy purposes. In some tools you can opt out of saving chat history, which will usually mean inputs will not be used for training. 
  • The guidance for use of Microsoft Copilot at UBC, for example, notes that starting a new chat will clear the chat history. 

1f. Copyright and intellectual property 

GenAI tools are often developed using large datasets, which might include copyrighted material. This could influence the AI’s outputs, occasionally resulting in the inclusion of copyrighted works. When utilizing GenAI outputs, be aware of the possibility that these may mirror existing copyrighted works, potentially leading to inadvertent infringement. 

  • Your intellectual property: 
    • Educators: Be aware about entering personal intellectual property such as lecture notes, syllabi, or other teaching materials into GenAI tools if you wish to retain control over their distribution and use.  
    • Students: Be aware about inputting your own academic work into GenAI tools, as this may lead to unintended sharing or use beyond your control. 
  • Use of third-Party intellectual property: 
    • Uploading third-party material that is someone else’s intellectual property, including journal articles, book chapters, faculty or TA teaching materials, or other copyrighted works, to GenAI tools may constitute copyright infringement. A safer approach is to only upload such materials with the express permission of the copyright owner, or if such use falls within the bounds of Fair Dealing.  
  • Terms of service compliance:  
    • Always review and comply with the terms of service of any GenAI tool you use. This is crucial to understand the permissible use of the outputs and any attribution requirements. 
    • Note that different GenAI tools have varying policies regarding the ownership of outputs. Some may grant users ownership of the outputs, while others might retain certain rights or impose more restrictive terms.  
  • Distribution and sharing of AI-generated materials: 
    • Exercise caution when distributing or sharing AI-generated materials, as they may incorporate or be derived from copyrighted works.