Artificial intelligence

GenAI is an umbrella term for a type of machine learning and a group of algorithms that can create new content, such as text, code, images, videos, and music, or a combination of all of these formats. GenAI generates output in response to a query or prompt using generative models such as Large Language Models (LLMs) that rely on large datasets.

Some well-known examples of genAI tools are text generators, such as ChatGPT, Bing Chat Enterprise, and Google Bard, and image generators, such as DALL-E and Midjourney. This list of examples is very small – there are many other tools that have incorporated artificial intelligence capabilities.

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 genAI system is a large language model that has been trained on a dataset to interact with users in a conversational way. ChatGPT is part of the broader category of genAI.

The current version of ChatGPT bases its output on its training data of internet content up to and including 2021.

Although ChatGPT is currently freely available, OpenAI has also launched paid versions via a subscription plan.

Microsoft made Bing Chat Enterprise available in preview on Bing.com and the Edge sidebar.  

Bing Chat Enterprise provides AI-powered chat with commercial data protection. Bing Chat is integrated with Microsoft Edge, allowing users to access the chatbot through the browser's sidebar. This integration ensures that the chatbot is accessible and can be used alongside other productivity tools, such as Microsoft 365 applications. Bing Chat Enterprise is grounded in web data and provides verifiable answers with citations (although sources may not have reliable or accurate information), along with visual answers that include graphs, charts and images.  

Microsoft has included Bing Chat in the Office365 suite, and it is now part of the University of Manitoba’s M365 A3 license, which is available to all students, staff, and faculty.  

The University of Manitoba, however, has not formally approved the use of specific genAI software, genAI tools within our M365 A3 license should be considered with the same caution that would be applied to other third-party applications that ingest personal data.  

Large Language Models (LLM) are trained to predict the next word in a sentence, given the text that has already been written. Early attempts at addressing this task (such as the next-word prediction on a smartphone keyboard) are only coherent within a few words, but as the sentence continues, these earlier systems quickly digress. A major innovation of LLMs is their ability to pay attention to words and phrases which were written much earlier in the text, allowing them to maintain context for much longer and in a sense remember the topic of conversation. This capacity is combined with a training phase that involves looking at billions of pages of text. As a result, these models and their underlying technology, are good at predicting what words are most likely to come next in a sentence, which results in generally coherent text.

One area where genAI tools sometimes struggle is in repeating facts or quotations. This means that the models may generate claims that sound real, but to an expert are clearly wrong.

GenAI may also struggle in discussion of any event or concept that has received relatively little attention in online discourse. To assess these limitations, you could try asking the system to generate your biography.  Unless there are numerous accurate biographies of yourself online, genAI tools are unlikely to generate a comprehensively correct biography.

At this time, ChatGPT and Bing Chat are free to access and to trial or preview. Instructors might be interested in assessing the capabilities and limitations of genAI tools. They may also wish to explore 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.

However, confidential information should never be entered into a genAI tool. All content entered may become part of the tool’s dataset and may inadvertently resurface in response to other prompts.

Yes. Some instructors may wish to use the technology to demonstrate how it can be used productively, or what its limitations are.

Keep in mind that asking or requiring your students to access these tools is complicated by the fact that genAI tools have not been vetted by the University of Manitoba for privacy or security. The broad or mandatory use of such systems by students for the purpose of learning or assessment are discouraged until we are assured that the system is protecting personal data that is entered into such systems (e.g., the email address used to register on the system).

If you decide to ask or encourage students to use any genAI systems in your courses, there are a few issues to consider before you do so:

• Never input confidential information into a genAI tool. All content entered may become part of the tool’s dataset and may inadvertently resurface in response to other prompts.

• There may be some students who are opposed to using AI tools. Instructors should consider offering alternative forms of assessment for those students who might object to using the tools, assuming that genAI is not a core part of the course.

• Instructors should consider indicating on their syllabus that genAI tools may be used in the course and, as relevant, identify restrictions to this usage in relation to learning outcomes and assessments.

• Be aware that not everything that genAI technology produces is correct. You may wish to experiment with the technology to see what kinds of errors it generates; citations are often fabricated, and inaccurate prompts are sometimes taken as fact.

• There is a risk that genAI may perpetuate biases inherent in the material on which it was trained.

• Users may also find that there are times when ChatGPT is unavailable due to high demand or that Bing Chat crashes. If you plan on using genAI tool ‘live’ in the course, consider having a back-up plan. 

• OpenAI or other genAI providers may also change the terms of use without notice.

Students are expected to complete assessments on their own, without any outside assistance, unless otherwise specified by their instructors. Instructors are strongly encouraged to speak to their students about what tools, if any, are permitted in completing assessments. Written assignment instructions should indicate what types of tools are permitted; vague references will generally not suffice. If adding a prohibition on genAI tools to assignment instructions, it is best to state that the ‘use of generative artificial intelligence (or genAI) tools’ is prohibited, as opposed to the use of one particular tool, as there are many generative AI tools available.

A document is available for download at the top of this webpage that contains sample language that instructors may include in their course syllabi to clarify for students if the use of genAI tools for completing course work is acceptable, or not, and why. Instructors are also encouraged to include information on specific assignment instructions to explicitly indicate whether the use of genAI is acceptable or not.

If an instructor indicates that use of genAI tools is not permitted on an assessment, and a student is later found to have used such a tool on the assessment, the instructor should consult the Student Discipline Bylaw, and/or the Department Head or Dean.

Some students may ask if they can create their assignment outline or draft using genAI, and then edit that draft. Before discussing the assignment with your students, consider what your response to this question might be, and address this question in advance.

Instructors may wish to consider tips for assessment design here, or meeting with, or attending a workshop at The Centre for the Advancement of Teaching and Learning to get more information about assignment design. Consider what the learning goals are for the assignment, and how students can best achieve those considering this new technology.

The University of Manitoba discourages the use of AI-detection software programs on student work. Such software programs are unreliable and often incorrectly flag instances of genAI use in human-written content (Weber-Wulff, 2023). Some of the AI-detection software programs assess if a piece of writing was generated by artificial intelligence simply on its level of sophistication. Using such detection programs could negatively impact students if they were to be improperly accused of using a genAI tool. 

Sharing your students’ work with these software programs without their permission also raises a range of privacy, intellectual property, and ethical concerns.

Instructors are encouraged to continue to use their traditional methods for detection of potential academic misconduct, including:

• lack of citations
• citations that do not match the content
• false citations
• topic of papers/essays veer off from assignment instructions
• content that is too general
• similarity to text generated using artificial intelligence using assessment instructions as a prompt
• argumentation that is too sophisticated for the level of the student
• sweeping statements that a student could not have generated on their own
• significant contrast between the writing and other samples of writing from the same student in the same course (e.g., assignment vs. final exam)
• a great essay from a student who rarely attended class.

Some of the methods described above require instructors to have a solid understanding of the level and style of their students' academic work.

Meeting with a student to discuss their assignment in person, as per the Student Discipline Bylaw – Academic Misconduct Procedures, is another option. If you have suspicions about a submitted assessment, engaging the student in a conversation about their assessment can be helpful to understand the issue. Refer to this UM Discovery Interview Tips resource for examples of questions you may ask to gain clarification.

Yes. If you use multiple-choice quizzes or tests, assume that genAI systems will answer the questions unless they pertain to highly specific subjects, new knowledge, or the specifics of classroom discussions, or any content that is unlikely to be found on the internet. Some instructors may wish to test this by using their multiple-choice or short answer assessments as prompts, and reviewing genAI's responses.

Talking to students about genAI and its limitations will let students know that you are well aware of the technology, and will likely generate interesting discussion and help to set guidelines for students. Let students know verbally and in assignment instructions what tools may or may not be used to complete the assignment. Advise students of the limitations of the technology, and its propensity to generate erroneous content.

If you choose not to allow use of genAI tools on your assignments, here are some tips for generating assignments to which genAI systems will have difficulty responding. Some include:

• ask students to respond to a specific reading, particularly one that is from the last year, and may not be on the internet, or may not have generated much commentary online. Generative systems struggle to create accurate responses to prompts for which there is little or no information on the internet.
• ask students to create a video or recording that explains or expands on their work. 
• use a flipped classroom approach, and/or assign group work to be completed in class, with each member contributing.
• ask students to create a first draft of an assignment, or an entire assignment, by hand in class. (Consider the accessibility needs of students who may require accommodations.)
• call on students in class to explain or justify elements of their work.
• ask students to use genAI to generate material, and then ask them to critique the responses.
• talk to your colleagues about ideas for your discipline. Different disciplines, such as computer science, history, language studies and visual studies may be developing new norms of pedagogy.

The University of Manitoba asks that you not submit student work to any third-party software system for grading, or any other purpose, unless the software is approved by the University of Manitoba. A completed assignment is the student’s intellectual property (IP), and should be treated with care.

The University of Manitoba currently has licensed software tools available for facilitating grading, such as Crowdmark. Please note that the instructor is ultimately responsible for ensuring the grade accurately reflects the quality of the student’s work, regardless of the tool used.