Transforming Learning: OpenAI’s impact in the classroom

OpenAI’s efforts to revolutionise various aspects of education using their large language models, including GPT-3, GPT-3.5, and GPT-4, have taken a new turn. In a recently released blog post titled Teaching with AI, the organisation showcases the experiences of six educators from different parts of the world, most of whom teach at the university level (though one teaches at the high school level). These educators have integrated the LLM into their classrooms.

OpenAI shares, “We’re presenting a few stories of how educators are using ChatGPT to enhance student learning and providing prompts to assist educators in getting started with the tool.”

The educators’ use of this AI tool in their classrooms varies. One educator employs the tool as an educational role player, taking on roles like a debate opponent or recruiter to engage students in discussions. Another teacher uses LLM to aid English-as-a-second-language students with translation assistance. Additionally, some educators have their students fact-check the information generated by this LLM.

OpenAI has also included sample prompts developed by AI influencer Ethan Mollick, a professor at the University of Pennsylvania Wharton School, and his wife, fellow professor Lilach Mollick. These prompts are designed to help teachers with lesson planning and can even transform the default ChatGPT into an ‘AI tutor’ for students.

Introducing generative AI into the classroom has not been without controversy. Some students have sought to use it as a means to bypass their coursework, such as writing essays. As a result, several educational institutes worldwide have banned ChatGPT and added it to their Internet network blocklists. Nevertheless, the New York City Public School system reversed its decision in May, allowing teachers to use LLM as they see fit.

OpenAI previously made headlines by releasing an ‘AI Text Classifier’ designed to identify AI-generated content, but it was discontinued due to its low accuracy rates. In a recent Educator FAQ, OpenAI delves deeper into the issues related to the Text Classifier.

In response to the question, “How can educators respond to students presenting AI-generated content as their own?,” OpenAI explained, “While some, including OpenAI, have released tools claiming to detect AI-generated content, none have proven reliable in distinguishing between AI-generated and human-generated content.” OpenAI’s attempt at training an AI-generated content detector misclassified human-written text, including well-known works like those of Shakespeare and the Declaration of Independence, as AI-generated.

The organisation also acknowledged that such tools might disproportionately impact students learning English as a second language or those who write in a formulaic or concise style. Furthermore, students can make minor edits to avoid detection.

The company stressed the importance of having a “human in the loop” when it comes to assessment. Some teachers have started asking students to share their interactions with ChatGPT as a way to demonstrate their critical thinking skills.

 While OpenAI recognises that its LLM can be a valuable tool for providing students with feedback, they caution against relying solely on these models for assessment decisions. The idea of replacing teachers’ or students’ active involvement in their coursework with the AI tool is not on the horizon, and OpenAI aim is to promote the LLM as a complementary tool for both educators and learners, alongside traditional classroom resources like pencils, notebooks, computers, and globes.