Thinking about ChatGPT?

ChatGPT (GPT stands for Generative Pre-trained Transformer) is a tool developed by OpenAI that is capable of producing human-like responses to prompts. This AI system is a large language model that has been trained on a dataset to interact with users in a conversational way. The current version of ChatGPT bases its output on its training data of internet content up to September 2021.   

While ChatGPT is widely known, it is part of the broader category of generative AI, a form of artificial intelligence that generates new content based on the data it has been trained on. The new content that is generated can be text, images, code, videos, etc. Other examples of generative artificial intelligence tools (“AI tools”) include Bing, Bard and Dalle-2. 
 
Some generative AI tools are available for free, while others such as ChatGPT Plus have a paid subscription version with further features. Some tools are connected and others offer plugins to connect to the internet and others are not connected and are limited to the information they are trained on which may not be up to the present day.    

Instructors are welcome to test AI tools to assess their capabilities and limitations and experiment with how the tools respond to particular course assignments or prompts. For example, by using prompts from their own assignments and assessments, instructors can gain a sense of how the tool could potentially be used by learners as well as its limitations. Keep in mind that some students may be sophisticated in prompt engineering and could be able to prime GenAI tools to return results that are superior to those obtained by entering an assignment question into a fresh chat.  Additionally, in trying out tools, be sure not to share in prompts any personal or sensitive information, or any information you wish to be kept private, as this content may become part of the dataset the models train on. 

The use of ChatGPT or other generative AI tools does not automatically equate to academic misconduct at UBC. At this time, whether the use of AI tools in courses is or is not allowed is a course or program-level decision and there is no overall, UBC-wide ban on its use in teaching and learning.  Individual instructors should clarify expectations with their students early in the term, such as on the syllabus. If instructors have questions about any Department of program level policies on artificial intelligence tools, they should reach out to their Department or program. Further information is available on generative AI syllabus language.  

• If using ChatGPT and/or generative AI tools on coursework has been prohibited by the instructor, then using these tools would be considered to be academic misconduct.   
• If using ChatGPT and/or generative AI tools has been permitted by the instructor, then instructors should make sure to convey the limitations of use and how it should be acknowledged and use should stay within those bounds.  
• If the use of ChatGPT and/or generative AI tools has not been discussed or specified by the instructor, then it is likely to be considered as prohibited as an example of the “use or facilitation of unauthorized means to complete an examination or coursework” and more specifically as “accessing websites or other online resources not specifically permitted by the instructor or examiner” (Discipline for Academic Misconduct, Vancouver and Okanagan 3.1.b.iv), and potentially plagiarism (3.1.e).  

 
Students should not assume that all available technologies are permitted. If students are not sure about whether AI tools are allowed, as with any tool, they must ask their instructor for clarity and guidance.    

UBC’s Academic Calendar provides guidance on what is considered academic misconduct (Academic Misconduct by UBC Students, Vancouver and Okanagan). Academic misconduct is any conduct by which a student gains or attempts to gain an unfair academic advantage or benefit, thereby compromising the integrity of the academic process. Artificial intelligence tools like ChatGPT are not expressly named in the academic misconduct regulation but their use could be considered an attempt to gain an unfair academic advantage, as well constituting unauthorized means to complete an assignment or assessment, the accessing of a website that is not permitted, or other, depending on the specific case.  

It is important for instructors to address academic integrity throughout the semester. Create opportunities to discuss expectations around academic integrity ahead of assignments and exams and throughout the term. Refer to academicintegrity.ubc.ca for tools to teach academic integrity and respond to academic misconduct. 
 

There are many innovative and creative ways that an instructor might use generative artificial intelligence technologies in their courses. From open conversations around the ethical and societal implications of artificial intelligence to dialogues with chatbots and critical engagement with their outputs, there is potential both for learning about and for learning to use the technology.   

If instructors choose to integrate ChatGPT or other GenAI tools into course activities, they should consider the privacy implications of doing so. Using ChatGPT, for example, requires a login that asks for personal information, and it is important to offer students an alternative option if they do not wish to provide this information to engage with the tool themselves. One option could be for the instructor to generate conversations and share them with students to engage with outside of the tool.  

Suggestions and examples around how generative AI can be integrated into teaching and learning, as well as how to design assignments and assessments that make it harder to use AI, have been developed by the CTLT.  Instructors may wish to review the resource from the CTLT Teaching and Learning in an Era of Generative AI or AI in Teaching and Learning. 

A key expectation of academic integrity for students is completing their own work. Besides producing essays in seconds, generative AI has proven itself capable of completing multiple-choice exams and short answer questions, generating code, and producing creative output.   

While artificial intelligence technologies should not be used to complete academic work, there may be times where instructors and students engage with it as part of student learning. If this is the case, instructors should provide clear guidance to students around how they are allowed to engage with the tools.   
 

UBC discourages the use of artificial intelligence detectors on student work, and is not at this time planning to purchase or support any such tools at the institutional level.  
 
There are several AI detectors currently in existence, such as GPTZero, Turnitin, and AI Content Detector. Despite the availability of such tools, it is important to remember that they might not be fully tested and that the technology to potentially outwit them continues to evolve.  The detectors are not foolproof and can produce false negatives and false positives. It may also be possible for the user to modify content to avoid detection.  Instructors might wish to consider UBC’s response to Turnitin’s AI detection feature and the concerns that were raised. 

If instructors still choose to use AI detectors, they should be aware and understand their limitations and issues they can raise. None of the detectors has undergone a UBC Privacy Impact Assessment, and as such there may be privacy and security concerns with submitting student work to them, particularly without their knowledge or consent. Instructors should not use these tools to evaluate any student work that contains the name of the student or any other personal information of the student or third parties. If student work may be submitted through one or more AI detectors, instructors should be transparent with students and let them know at the beginning of the term, such as in the syllabus.  

Finally, it is recommended that AI detection tools not be used as the sole factor in decision-making around an allegation of academic misconduct. If an instructor suspects that an assignment or assessment has been completed with unauthorized use of AI tools, they should proceed as they would for any other potential allegation of academic misconduct. An overview of the academic misconduct process for instructors is available on the academic integrity website. 

No, currently Turnitin’s new AI-detection functionality is not available for use in any UBC course. UBC has reaffirmed their decision to not enable Turnitin’s new AI-detection feature. 

On April 4 2023, Turnitin activated a new and separate feature that attempts to identify AI-generated text . The LT Hub Leadership group, with the support of the Provosts at both UBC Vancouver and UBC Okanagan, made the decision not to enable this feature at that time (April 2023) and has recently reaffirmed their decision (August 2023) for the following reasons:  
 

Effectiveness of the feature is still unclear  

• Testing for accuracy in the AI-detection feature is in early stages. 
• Testing for potential bias in the feature continues to be in early stages. 
• Ability of the feature to keep up with rapidly evolving AI is unknown. 

It is not possible to double-check or review the results  

• Instructors cannot double-check the feature results.  
• Results from the feature are not available for students to review.  

You can read more about UBC’s decision to not enable Turnitin’s new AI-detection feature on UBC’s Learning Technology Hub website.  

UBC is continuing to wait before deploying any AI-detection features, including the one by Turnitin, as more information is needed about the feature’s effectiveness, accuracy and bias mitigation, and about the ability to employ an AI detector as a robust component of talking to students about suspected academic misconduct. The use of other AI detection tools is also not recommended, due to similar issues as those noted above, as well as privacy and security concerns; to date, no AI detection tool has undergone a UBC Privacy Impact Assessment process.  

Instructors who suspect that a student has used generative artificial intelligence tools contrary to expectations should follow the standard academic misconduct process.  If an instructor has a suspicion based on the student’s work, they should follow the procedure as they would for any misconduct allegation. Instructors should not rely on AI detectors to form the basis of an allegation of academic misconduct . If students have concerns about any allegations of academic misconduct against them, they can reach out to the Office of the Ombudsperson for Students, AMS Advocacy or SUO Advocacy.

There has been much conversation about how generative AI should be cited if its use is permitted in assignments and academic publications. The American Psychological Association (APA), Modern Language Association (MLA) and the Chicago Manual of Style have all provided recommendations in this area. Further information on this topic can be found in the Generative AI and ChatGPT LibGuide from the UBC Library. 

 Funding is available for projects that explore the intersection of generative artificial intelligence and teaching and learning, as well as other teaching and learning topics (Teaching and Learning Enhancement Fund, Aspire 2040 Learning Transformations Fund, SOTL Seed Program). Explore teaching and learning topics and further support through the CTL and the CTLT.