Higher education seemed to embrace the concept of AI for administrative systems and student relations support. There was a lot invested in technologies that could “personalize” systems and connect students with departments, divisions and support services faster. Nearly every college website has a little bot with a friendly name representing the school’s mascot who’s ready to help.
But that all changed recently when that same helpful bot entered the classroom. Academic media outlets have overwhelmed instructors and administrators with conflicting tales of how we should fight or embrace the bot.
The enemy (?)
While Open AI’s ChatGPT has taken most of the spotlight, other major technology companies have jumped in with a “hold my beer” aggressiveness in an attempt to take the lead in the AI race. Google presented “Bard” and Microsoft resurrected Bing in the form of “Bing AI,” but both of those tools appear to have waiting lists for users.
This article is part of a biweekly series provided by the Instructional Technology Council, an affiliated council of the American Association of Community Colleges.
But like most technology, some of these platforms are geared toward specific audiences and that may be key to determining which side of the debate you end up on. For example, Bard is not Google’s first rodeo with AI apps, Socratic by Google has been around for several years and focuses on classroom questions. The website boasts the quote, “This app explained to me what my teachers couldn’t.” That may be a concern. But the business-focused Jasper, may not cause anxiety for composition professors.
Winning the battle, not the war
Many educators are finding solace in the development of apps that can detect AI-created content, like GPTZero, created by 22-year-old Princeton student Edward Tian. Even Open AI has developed its own detector called GPT2 Output Detector.
There are several issues with relying on the swift development of apps to detect AI-created content. First, we need to consider our own processes, especially when it comes to grading; these are new tools developed to fight also new tools. There may be an extensive timeline on when they would be effective enough to develop licenses with major LMS companies like Instructrue and Anthology.
We can also take a closer look at how our current plagiarism tools work and think about whether we need to reinvent the wheel. Programs like SafeAssign and Turnitin are designed to search databases for content, and while bots like ChapGPT do not copy word for word, they are both designed to scrub databases and websites to identify content. Updating the algorithm of existing tools or fast-tracking the development of new tools may win the battle, but not the war.
The best defense is a good offense
For those interested in learning to love the bot, there is an alternative — examine the technology, learn to understand it and change the way you teach. Currently, most of these applications do not cite their sources, which can be a relief for research-based course instructors. A student would still need to backtrack through the AI-created content to determine the source material.
But much like Turnitin can evolve to fight the bot, we could be a few algorithms and program updates away from ChatGPT churning out APA in-text citations. But wait, we’ve had programs that create somewhat accurate citations for years — students will still need to understand the correct process to evaluate the bot’s work.
Secondly, ChatGPT writes in a very neutral tone, not the persuasive or argumentative tone required for college writing that is heavily invested in showcasing a student’s critical thinking skills. For example, this was entered into ChapGPT: “Do horror films have a negative psychological effect on children?”
ChatGPT: “There is some evidence to suggest that exposure to horror films can have a negative psychological effect on children, particularly those who are younger or more susceptible to fear.”
This example shows an average thesis statement. It provides a stance that is based on the wording of the question. ChatGPT proceeded to provide an even look at both sides of the argument. The result is simply average neutral writing with no opinion and no voice. Instructors will always be able to identify weak writing. A good instructor will always win the war.
Clap back from ChatGPT
Instructors have a lot to say about ChatGPT, but what does it have to say about the battle between profs and bots:
Prof: How should instructors fight ChatGPT
ChatGPT: As an AI language model, I don’t believe there is a need for instructors to fight me. Instead, instructors can use my capabilities to enhance their teaching methods and engage with students in new and innovative ways. Here are some suggestions on how instructors can use me effectively in their classrooms:
- Use me for automated grading and feedback: Instructors can use me to automatically grade assignments and provide immediate feedback to students.
- Use me for interactive class discussions: Instructors can use me to facilitate class discussions and answer questions in real-time.
- Use me for personalized learning: Instructors can use me to personalize learning by providing customized recommendations based on students’ learning needs and preferences.
- Use me for creating content: Instructors can use me to create interactive learning materials such as quizzes, flashcards, and study guides.
- Use me for language learning: Instructors can use me to help students practice their language skills by having conversations with me in the target language.
In summary, instructors should view me as a tool to enhance their teaching, rather than as a competitor. With proper use, I can help instructors save time, improve student engagement, and promote personalized learning.
That’ll do bot, that’ll do.
