Pedagogical Implications of Chatting with AI to Learn

Ross Jackson, Director, Masters of Science in Analytics, Associate Professor, Wittenberg University

Ross Jackson, Director, Masters of Science in Analytics, Associate Professor, Wittenberg University

Inquiry-based learning extends back to at least the time of Socrates. Instructors asking questions designed to encourage student reflection and critical thinking resides at the core of both the Socratic method and constructivist learning approaches to education. The recent development, and rapid expansion of generative artificial intelligence (AI) technology, holds potential for radically altering long-established norms in education. There are pedagogical implications of students “chatting” with AI to learn, rather than interacting with instructors and classmates directly. As with any significant shift in paradigms, there are both opportunities and risks. Adaptive educators will work to guide student use of AI to enhance its potential and mitigate its liabilities.    

No matter how laidback one is, there is always a perceived and real gap in power, authority, and relatedness between instructor and student. In most traditional approaches to education, the teacher oversees content and assignments, and the students are expected to follow directions and perform. This is perhaps most obvious in terms of the topics selected for focus and the rate and depth of coverage. AI increases student autonomy in learning. Whereas this does not necessarily or completely change the structural basis of authority between teacher and student, it certainly does change the dynamics of the educational exchange. By using generative AI technologies students can explore a topic to their individualized desired depth of inquiry, and then pivot among and between interrelated topics. Through the process, there is potential for students to gain a more holistic understanding of topics, and to narrow their focus of inquiry to subfields of interest. AI enables students to tailor their education in a way that would simply not be practicable under more traditional approaches. Additionally, AI removes one of the most limiting factors in learning, fear.

“This group consensus ensures that student protection, including the oversight of student data, receives continuous support”

Understandably, students can be self-conscious and uncertain in their learning. Publicly discussing something new can be inherently uncomfortable. It is even more so when one is sensitive to being “wrong.” This can be a significant barrier to student engagement and learning. When students feel secure, it is relatively easy to gain their active participation in learning-focused discussions. However, many students struggle to achieve that level of comfort in the classroom. The traditional approach to education requires instructors to either call on reluctant students to participate, or simply hope that some of the benefits derived through the discussion cascade to those who are merely watching the conversation unfold rather than engaging in it directly. AI provides a basis for students to participate in a discussion, albeit virtually, without fear or timidity. There is potential here. Because students can “ask” generative AI things they might be reluctant to ask a teacher in a classroom discussion, students can get answers to their deeper questions. These benefits of AI are worthy of pursuing. However, the adoption of AI as a pedagogical tool, is not without significant risk.

AI didn’t create itself, and its operation and maintenance isn’t free, even when there are no explicit costs associated with using a given AI tool at the point of employment. Educators benefit from knowing who owns the AI tool being used in the classroom, why the tool was created, the type of data it was trained on, how interface data are stored and subsequently used, as well as any limits in terms of responses allowed. This information will likely be both time consuming and difficult for instructors to obtain.

It is not uncommon for politicians to decry the “liberal bias” of education. Since ideology is ever present, it is ultimately unavoidable. There will always be ideology. It is simply a matter if one understands what ideology is dominant. The expansion of AI will replace the current ideology, whatever its content, with one that is more opaque. There is danger associated with private companies, with undisclosed motives, creating and providing the AI tools that will be used to inform and educate society. Paradoxically, the skills needed to assess the desirability of incorporating AI in education are the ones potentially at risk of atrophying under its use.  

Educated members of our society need to be able to engage meaningfully in critical thinking. The consequences of a deficit of this skill are increasingly visible in our society. It is certainly possible that AI systems can be created to facilitate its development. It is unclear if those who are committed to that principle will be the ones building the AI tools which are used. Corporations have an incentive to develop AI tools that entice and entrap people into perpetual, subscription-based access for them to think.

AI-based learning can help students overcome the fear of being wrong, as it enables them to pursue unique topical development based on personal interest. This could be a transformative and empowering step forward in education. At the same time, the companies developing AI systems will likely obfuscate its operative ideologies and make students lifelong consumers of an access-based service which will become increasingly necessary for them to be able to think. The pedagogical implication of chatting with AI to learn is that it will simultaneously enhance and constrain freedom and knowledge, and it requires fluency with and detachment from AI to assess critically its quality. 

AI