“The traditional college essay is dead,” were the words of Stephen Marche[1] about how the traditional/conventional teaching methods will be “disrupted from the ground up” by ChatGPT. As Marche (and many other writers[2] [3]) suggests, ChatGPT has the potential to revolutionise the way education is delivered and assessments are conducted in business schools and other institutions of higher education. In this article, we will take a closer look at the pros and cons of the emergence of Large Language Models, such as, ChatGPT and its impact on business schools in the higher education sector. We also discuss potential implications of this technology.
What is ChatGPT and how does it work?
ChatGPT is a powerful language model developed by OpenAI[4]. It is based on the GPT (Generative Pre-trained Transformer) architecture, which uses deep learning techniques to generate human-like text. One of the key features of ChatGPT is its ability to generate high-quality written content in a matter of seconds (subject to the complexity of the input). It can be used to generate unique and creative ideas (e.g., jokes for an open mic night[5], fun facts[6] and even help to develop computer programming codes). In addition to its ability to generate written content, ChatGPT can also be used for other language-related tasks such as language translation, text summarisation, and question answering, in a number of languages. This makes it a versatile tool that can be used in a variety of fields and industries, including business education.
It is trained on a massive amount of data from the internet, allowing it to understand and respond to a wide range of topics and questions. The dataset used for training was from a comprehensive scan of the internet during the summer of 2021.
Why are educators interested (or concerned)?
ChatGPT has sparked considerable interest (and raised concerns) amongst educators in higher education because as it has the potential to disrupt higher education, especially in the area of assessment. One of the main concerns is that the technology could discourage students from learning basic concepts and skills. For example, if students can easily generate high-quality written content using ChatGPT, they may not see the need to spend time learning and practicing writing skills. In addition, if students can use ChatGPT to find information quickly, they may not see the need to develop research skills.
Another concern is that ChatGPT could be used to enable cheating in assessments/coursework. The ability to generate high-quality written content[7] in a matter of seconds could make it too easy for students to cheat in coursework. This is why it is crucial for educators and educational institutions to be aware of these potential risks and to implement guidelines and best practices to prevent any misuse or cheating. This includes, for example, using plagiarism detection software (e.g., RoBerta[8]) and educating students about academic integrity.
Moreover, some educators are concerned that the use of ChatGPT could lead to a loss of authenticity in the assessment process. For example, if students are able to use ChatGPT to generate assessment answers, it may be difficult to determine the students’ true understanding of the material. This could lead to a situation where students are passing assessments without achieving the intended learning objectives.
What can it do?
One of the most notable advantages of ChatGPT is its ability to generate textual content very quickly. For example, entering “Write an essay about the impact of Global Warming on businesses” into ChatGPT returns the following response (refer to Figure 1). This can be incredibly useful in the context of business schools, where students are often required to submit large amounts of written work, such as essays, reports, and dissertations. With ChatGPT, students can quickly and easily generate answers that are tailored to the specific requirements of their assignment. Although this can help to save them time and reduce the stress associated with completing large amounts of written work, it also has the potential for the students to not achieve the learning outcomes/objectives.
Figure 1: Response from ChatGPT
Does it have limitations?
ChatGPT is not without its limitations. For example, it can struggle with understanding the nuances of human language and may not be able to fully grasp the context of a particular assignment. The example in Figure 2 found online (no reference) clearly shows a limitation of ChatGPT in this regard. The model is trained on a massive amount of data from the internet, but it may not always fully understand the context of a particular assignment. This means that it may not be able to generate responses that are completely accurate or relevant to the task at hand. Further, ChatGPT does not have the ability to understand emotions and social cues, and this is a limitation that could affect its ability to understand the context of a particular task.
Figure 2: Response to a prompt (Source: online no reference)
Another limitation of ChatGPT is that it is trained on a finite dataset (as at the time of writing this blog), which means that it is only able to understand and generate text based on the data it has been trained on. The training data of ChatGPT is usually up to a certain date, in this case, the data was trained up until summer of 2021, so it may not be able to understand or generate text about events, trends or information that have happened after that date. For example, if you were to ask ChatGPT about the potential reason for the resignation of New Zealand’s ex-Prime Minister (Jacinta Arden), you will not be given any answer, as it would not be able to include information about the events that have occurred after the summer of 2021. As we know, the pandemic is ongoing and it has had a significant impact on the global economy, which is not captured in the data that ChatGPT was trained on. This would limit the tool’s ability to provide an accurate and up-to-date analysis of the topic.
Closing Remarks
The implications of the emergence of AI models, such as ChatGPT on business schools and the wider HE sector is a complex issue with valid arguments on both sides. We cannot ignore the existence of ChatGPT (and similar tools, for example, Google BARD, OpenAI DALLE), nor can we ignore that fact that students may use these tools for completing assessments. We also cannot ignore the difficulty of identifying its use in coursework and assessments in HE, thereby negating the effect of potential punitive measures (e.g., university policies) for its use in completing assessments. What we can, however, do as academics is to adapt to its existence by switching to assessments that take advantage of the limitations of ChatGPT. For example, increasing emphasis on critical analysis, more recent case studies, practical assessment questions, etc. If appropriate, it may be worth considering a change in current assessment modes to in-person, timed or oral examinations. The impact and implications of ChatGPT and similar AI-driven large language models in HE is a debate worth having, and it is time for educators, students, and institutions to engage in this conversation to explore how this can impact HE.
[1] https://www.theatlantic.com/technology/archive/2022/12/chatgpt-ai-writing-college-student-essays/672371/
[2] https://www.timeshighereducation.com/campus/chatgpt-and-rise-ai-writers-how-should-higher-education-respond
[3] https://www.nytimes.com/2023/01/16/technology/chatgpt-artificial-intelligence-universities.html
[5] https://medium.com/illumination/14-creative-ways-to-use-chatgpt-you-probably-didnt-know-about-f0f5d2608c6d
[6] https://www.makeuseof.com/creative-ways-to-use-chatgpt-openai/
[7] https://www.edweek.org/technology/what-is-chatgpt-and-how-is-it-used-in-education/2023/01