ChatGPT and Generative AI Guidelines for Addressing Academic Integrity and Augmenting Pre-Existing Chatbots

Abstract

Chat Generative Pretrained Transformer (Chat-GPT) and related Generative AI models are leading a paradigm shift in the acceptance and application of Artificial Intelligence (AI) across all disciplines and industry sectors. Despite the criticisms of an ‘intelligence without knowledge or reasoning or the notions of truth’, ChatGPT is highly effective at human-like conversation with seemingly sophisticated and useful responses to questions, summarization, classification, extraction and generation tasks. Unlike similar large AI models in the modalities of image, audio and video, text-based conversation is straightforward and familiar to a large audience of regular users of the Internet and smartphone applications. This is further accentuated by the large-scale adoption of ‘standard’ chatbot technologies for trivial conversations in task-specific automation, across every industry sector. This rare combination of highly effective human-like conversation, familiarity of foundational technology and versatility of intelligent application, has led to several challenges and opportunities in leveraging generative AI. A primary challenge is its impact on the academic integrity of scholarly work, where AI-generated content can be useful and detrimental in both teaching and research. On the other hand, ChatGPT presents a unique opportunity in augmenting preexisting (‘standard’) chatbots with human-like conversation for advanced intelligent automation, across all application domains. Although diametrically opposed, the challenge of addressing academic integrity and the opportunity of augmenting pre-existing chatbots are grounded in the conversational AI capabilities of ChatGPT and similar generative AI models. In this paper, we investigate these formative capabilities and present guidelines for leveraging ChatGPT and similar generative AI models.