Historical GPTOnline editorial illustration for the September 2023 discussion of AI in education; it is not a classroom photograph or evidence of learning outcomes.
In September 2023, as a new academic year began, educators and students were navigating an era where artificial intelligence (AI) had become a critical part of the educational landscape. Conversational AI like ChatGPT, despite being in its developmental phase, had introduced new ways to engage students and alter teaching methods. However, these AI chatbots often blur the line between fact and fiction, raising concerns about academic integrity.
Higher education institutions were reconsidering how to handle AI tools, with approaches ranging from restrictions to classroom experiments. Policies differed by institution and course. Educators should explain permitted uses, attribution requirements and assessment expectations rather than assume that one rule applies everywhere.
OpenAI, the organization behind ChatGPT, published its “Teaching with AI” guide on August 31, 2023, which provides strategies for educators navigating AI in their classrooms. The guide explores ChatGPT’s capabilities, limitations, and potential applications, offering educators a fresh perspective on how to leverage AI.
AI tools can help educators shift focus from rote memorization to critical thinking. For example, educators can ask students to evaluate a chatbot’s output and explain and correct its errors using course sources. However, it is important for educators to ensure that AI-generated content is accurately represented and not mistaken for absolute truth.
The issue of academic dishonesty involving AI tools prompted interest in detectors such as GPTZero and Turnitin. These tools estimate whether text may be AI-generated; they do not establish authorship or misconduct. False positives can flag legitimate student work, while edited AI-generated text may be missed. Turnitin’s guidance warns against using its result as the sole basis for adverse action against a student. Review the work, drafts and relevant course rules with human judgment.
Rather than treating a detector score as proof or assuming that every assignment should allow AI, match the policy to the learning objective. Where AI is permitted, students can record their prompts, explain which suggestions they used and show how they checked the result. Avoid putting identifiable student work or records into tools that the institution has not approved for that purpose.
ChatGPT’s ability to handle diverse prompts allows educators to tailor AI interactions to their specific needs. Whether it is answering student questions, generating discussion points, or helping to rephrase a document for a particular audience, AI tools can assist with parts of course preparation, although educators still need to check accuracy and suitability.
OpenAI’s guide includes examples from Ethan Mollick and Lilach Mollick, identified in the 2023 guide as being at Wharton Interactive, demonstrating how AI can be used to create lesson plans, draft explanations, and even serve as an AI tutor. These examples offer a starting point for educators looking to more deeply integrate AI into their teaching.
However, the inherent limitations and biases of AI must not be overlooked. Generative AI tools, trained on extensive datasets, can propagate Western biases and reinforce certain stereotypes. OpenAI’s guide provides advice to educators on how to help students mitigate these biases, encouraging more socially responsible use of technology.
There are various educational resources available to help instructors get comfortable with AI in the classroom, including Auburn University’s “Teaching with Artificial Intelligence” course. Check the provider’s current access and enrollment terms.
The 2023 discussion highlighted both useful teaching experiments and unresolved risks. Educators can test a small, clearly defined activity, review the model’s output and ask students to demonstrate their own understanding before expanding its use.
