As we head into another school year, many professors and educators are grappling with a difficult question: What the #%**@ do we do about ChatGPT?
Way back in the last century, when Google first appeared (1998, actually), we remember debates among our academic colleagues as to whether students should be allowed to use a search engine to gather research for assignments.
It’s hard to believe professors thought like that 👨🎓!
There were similar conversations around Wikipedia. And even how to shut down social media in the marketing classroom.
Of course, no one would have that attitude toward Google, Wikipedia, or social media today. They are essential tools students can use throughout their education.
We believe academic acceptance of ChatGPT and other AI tools will follow a similar trajectory.
Within the next couple of years, we believe students and professors will find it inconceivable not to use Generative AI in an assignment.
Human vs. Machine
Some academics are currently loathe to admit that ChatGPT and other AI writing and design apps are so adept at imitating human output, that it’s difficult to tell whether or not a human created a piece of written content or a visual.
But the reality is, if you ask a student to produce an essay, generative AI can turn out a pretty decent B paper (or higher if a student is adept at prompting and incorporates their own perspective).
That means many standard assignments, such as a 1,000-word essay on a given marketing topic, are no longer a good measure of learning because students can use AI to perform a serviceable job.
Further, ChatGPT detectors are known for producing false positive results, which makes it problematic for professors to check academic integrity.
Submitting to the AI Assignment
It’s incumbent on instructors to design new assessments that better measure whether or not a student has achieved the learning outcomes set out in a particular course. At the Schulich School of Business – York University, where we are both professors, we have worked throughout the summer to adapt our marketing courses for the upcoming year.
We have come to the realization that the use of generative AI tools and technology must be woven—not only into each individual course—but also into individual assignments.
That is, permissible usage for ChatGPT will vary depending on the student learning outcomes professors want to achieve and how easily an answer can be generated by generative AI.
For example, in the marketing research course David teaches, students are asked to write a survey for an organization. In the syllabus, David prohibits students from using ChatGPT to write the survey. Students simply would not learn very much by having a machine learning algorithm crank out a questionnaire. However, once they finish the survey, David tells them it’s OK; indeed it’s optimal, to ask ChatGPT to ‘check the survey for errors in grammar, syntax and spelling’.
We believe this will improve the overall quality of the survey that students hand in for grading. Secondly, students are encouraged to use ChatGPT as a coach/mentor to review the finished survey and highlight sources of bias and areas for improvement. In this way, ChatGPT enhances student learning by giving them additional perspectives, perhaps not taught in the classroom.
Use at Your Own Risk
While students are allowed to use ChatGPT for components of their various assignments, we explain the risks of using Generative AI for their final output.
Because ChatGPT and other natural language generation tools have been trained on thousands, or even tens of thousands of articles, surveys, research papers, books, and so on, scraped from the Internet, a majority of the output one finds is mediocre at best.
So, while ChatGPT can offer suggestions for improvements, we don’t consider ChatGPT an expert in any domain including survey design, writing, editing, and so on. That’s why in David’s class, the final choice of wording for a survey question, and the responsibility for correctness, must fall to the student.
In Martin’s Digital Marketing Strategy course, AI is being incorporated into both content marketing assignments and in-class exercises. In one assignment, groups of students will be given a creative challenge and they will have to develop a full strategy with examples of the creative campaign and how they’d measure success in real-time.
In all cases, when students use Generative AI, they will be asked to submit both the machine-created content and a discussion of the prompts they used, how they evolved, and an analysis of what worked and why.
Students will also need to provide a rationale for how they fact-checked and verified AI-suggested references and how they would disclose their use of generative AI to the audiences they’re trying to reach.
Teaching the Future of Marketing Course
A good place to start is Wharton professor Ethan Mollick’s One Useful Thing newsletter, where he offers examples, strategies, and tips. You can also review OpenAI’s Teaching with AI guide to using AI strategically in the classroom.
In a new ‘Future of Marketing’ course we’re developing at the Schulich School of Business, we’re going all in on using ChatGPT, or similar natural language generators. Using AI is a basic requirement of all assignments/projects in the course.
In the course, students have the opportunity to develop full marketing campaigns using generative AI. This includes the exploration and use of text-to-text (ChatGPT, Claude, Bard), text-to-image (DALL-E, Midjourney and Stable Diffusion, text-to-video (Runway ML), text-to PowerPoint, and others. Students also explore the use of blockchain, augmented/virtual reality, and the arrival of Web 3 in a marketing environment.
In one of our first classes, we will invite students to explore generative AI capabilities in an in-class thinking exercise. We divide the class into two halves and ask them to brainstorm ideas to develop new ways to collect marketing research data from smart devices.
One half of the class will use ChatGPT, and the other half must rely on their own grey matter, writing their thoughts with pen and paper. And then we’ll ask the class to step back, analyze, and evaluate the results.
The generative genie is out of the proverbial bottle 🧞.
How well will this work? What will the student takeaways be? Will the students with access to ChatGPT outperform those who simply have to think of the answer on their own?
Right now, we don’t know, but we’re certain it will be a thought-provoking experiment and discussion for both the students and the professors.
The point is, we believe students must learn to explore the boundaries of technology.
And we will encourage a transparent, human-AI collaboration as a way to spark student’s thirst for knowledge, understanding, and creativity.
How are you adopting generative AI tools into your marketing courses and assignments? Please share your thoughts in the comments!
Connect with FMI
We would love to connect with you and hear your thoughts. You can reach us through our Website, LinkedIn, or Twitter. If you enjoyed this newsletter, please share it and subscribe to receive it directly in your inbox.