In our October edition of Edubytes, our guest editors are the CTLT’s Acting Academic Director Elisa Baniassad and Senior Educational Consultant Lucas Wright, to share how the CTLT is helping instructors navigating the era of generative AI.
Why hundreds of faculty members have attended the CTLT’s Generative AI Workshops and drop-ins, and why you should too!
Generative AI (GenAI) has changed (indeed it has bulldozed, quaked, and erupted) the landscape of teaching and learning. Large Language Models (LLMs) have moved at a breathtaking pace from being an unlikely to materialize and somewhat niche tool, through a pernicious challenge for academic integrity, to an exciting space of opportunity and creativity. They are evolving and gaining ground as we speak, creating seismic disruptions that will impact both how we teach our disciplines, as well as the nature of the disciplines themselves.
To ride this tsunami of change, hundreds of faculty members have attended CTLT GenAI workshops and visited the ai.ctlt.ubc.ca website to better understand the risks (technical, pedagogical, ethical, and even legal), how to mitigate them, and what the long-term challenges might be. The CTLT’s first workshop, a promptathon, saw hundreds sign up and stay tuned in for two hours. The monthly 30+30 Teaching with GenAI workshop series has consistently drawn an audience of 50 to 100 faculty members. The first of our weekly GenAI Drop-in clinics had over 30 faculty registrants in under 24 hours.
The major initial and enduring concern for many faculty members is the impact on academic integrity: how do we assess students when they have access to an always-on and mostly (and increasingly) correct tutor/editor/programmer/writer? Educators are beginning to draw borders around GenAI use, and some are promoting exploration while encouraging responsibility and safety. Time will tell if these guides work as the guardrails we hope.
But questions of assessment design—juxtaposed against the sheer scale of what GenAI can offer—are sparking deeper existential exploration in our disciplines: What do we really need students to know? What do we need to know that they know? What will they need to be able to do? What will they no longer need to do? In service of these investigations of purpose, our upcoming Teaching and Learning with Generative AI symposium will span the practical and the philosophical. It will provide a central, interdisciplinary forum for the pivotal and vital conversation not just about what we need to do about GenAI in teaching right now, but about how our practice, and even our mandate as educators, will fundamentally transform in the near, mid, and (decreasingly) far future.
How is GenAI Impacting Assessments and Assignments at the University?
Since the introduction of ChatGPT 3.5 last year, there has been a notable progression in its capabilities. In recent months, GenAI has enhanced its ability to successfully tackle university-level assessments across various disciplines. ChatGPT-4 has achieved scores in the top percentile on standardized exams across multiple subjects.
It has also produced critical reflections that not only consistently outscore student submissions but also remain indistinguishable to evaluators. In Science, Technology, Engineering, and Mathematics fields, and especially in Computer Science, GenAI has demonstrated proficiency in solving coding challenges. These developments have raised more questions than answers about the role of assessments in universities. They prompt a reevaluation of how faculty might adjust their current course evaluations to maintain academic integrity and rigor.
This is a complex challenge without clear answers. However, there are three approaches you can consider now to mitigate the challenge and to help you to determine when and how to appropriately integrate GenAI in your course design.
- Kick the tires: As recommended by Derek Bruff, a mathematics professor at Vanderbilt University, give this a shot. Take one of your course assignments and input it into a generative AI tool, such as Bing Chat or ChatGPT. Afterward, reflect on the questions provided to help you reassess the assignment.
- Invigilate in the classroom when appropriate: The advent of these tools has prompted some educators to contemplate the occasions when it's beneficial to conduct assessments or certain assessment elements in a face-to-face classroom setting. This could involve supervised exams or student presentations. Additionally, you could explore how active learning can be assessed in the classroom using methods like Peer Instruction and Team-based Learning.
- Emphasize the learning process: By emphasizing and charting the learning process, you can capture students' developmental stages. Activities should bring out their cognitive progression, and evaluations should focus on depth of thought, particularly where AI falls short.
Where is this going and how can we prepare?
Differentiating learning in higher education
GenAI technology can offer learners a more personalized educational experience. By mentoring students across various subjects, GenAI can deliver learning experiences that align with a student's proficiency level and cater to their individual preferences and approaches. Additionally, it can be adapted for specific subjects and knowledge areas. Achieving learner differentiation on a large scale has consistently been a challenge in university education.
On-demand learning resources
GenAI has demonstrated its efficacy in designing learning resources. For educators, this includes the creation of case studies, questions, and learning activities. For students, it offers the potential to develop tailored worksheets and activities for various topics. Given these advancements, how might we reimagine the classroom and teaching experience when such resources are readily available?
Regardless of the trajectory of these advancements, universities must persistently collaborate across diverse disciplines and fields. By doing so, we ensure that the human element remains central to the process and that these tools are utilized in both innovative and equitably fair ways in the realm of higher education teaching and learning.
Frequently Asked Questions on GenAI in Teaching and Learning
Below are three of the most frequently asked questions heard during the weekly GenAI Drop-In clinics hosted at the CTLT. Do you have a question of your own? Attend a session or access the recordings.
Intellectual property: Can GenAI tools like ChatGPT store a user’s input for other purposes?
The simple answer is “probably, yes.” Most GenAI tools indicate that any content submitted by users can be stored and used for the purpose of training their models. As a result, instructors should avoid submitting student-created content into these tools. Responses from GenAI models may also contain content that was not intended to be shared widely. Hence, caution is also needed when using what GenAI tools answer.
Academic integrity: Can I identify generated content?
The question of whether students are turning original work, or even work authored by them, is a classic academic integrity question. Students have always been able to plagiarize from other sources, but educatorscould at least search for the wording to see if it appeared elsewhere on the Internet or in literature. Tools like Turnitin could even help identify where students copied from. However, with GenAI, the landscape for academic misconduct is very different. The best resource for facilitating preparedness is the UBC Academic Integrity website, including resources on Academic Integrity and ChatGPT specifically.
Can GenAI help with course design?
Technically, GenAI models are capable of generating course content. There are first-hand accounts of using it for this purpose, including generating course outlines. From an ethical standpoint, there is an ongoing debate about using it, due to the risk of unknowingly using content form an online source and violating the copyright of the original author. Some educators are choosing GenAI as an ideation tool for their course’s structure. In that instance, a good practice is to disclose its use to students.
Sharing experience with AI in teaching and learning: The GenAI symposium
November 14-15, 2023
In this symposium open to the UBC community, you will hear from faculty, students, and staff at UBC about the significance of GenAI and how it is influencing teaching and learning. Additionally, you will gain insights into specific strategies for designing assignments and learning activities that incorporate and harness the power of GenAI. The event will feature a faculty roundtable, a student panel discussing the effects of GenAI on their educational experience, and more.
Register
Going further with prompts
Prompt engineering is the process of refining prompts that a person can input in a GenAI tool to generate text or images. It is one of the basic building blocks to transforming teaching and learning using GenAI. Below are resources to learn more about prompt engineering.
- A prompt pattern catalog to enhance prompt engineering with ChatGPT (PDF): Describes a catalogue of prompt engineering techniques presented in pattern form that have been applied to solve common problems when conversing with LLMs.
- Assigning AI: Seven approaches for students with prompts (PDF): Examines the transformative role of Large Language Models (LLMs) in education and their potential as learning tools, despite their inherent risks and limitations.
- MOOC: Prompt Engineering for ChatGPT: This course introduces students to the patterns and approaches for writing effective prompts for large language models.
Enjoyed reading about how GenAI is disrupting teaching and learning? Learn about other topics we covered in the October 2023 edition by reading the complete Edubytes newsletter. To view past issues, visit the Edubytes archive.
Are you interested in staying up to date on the latest trends in teaching and learning in higher education? Sign up for our newsletter and get this content delivered to your inbox once a month.
