Generative AI — Ana-Paula Correia

BY ANA-PAULA CORREIA

ChatGPT, also known as Generative Pre-trained Transformer, is a language model developed by the San Francisco-based company, OpenAI. The company was founded in 2015 by Elon Musk, Sam Altman, Greg Brockman, Ilya Sutskever, Wojciech Zaremba, and John Schulman commerce (OpenAI et al., 2015). The company is also responsible for creating other software programs such as GPT-3 and DALL-E2. These intensify the ongoing advancements in the field of Artificial Intelligence (AI).

Made publicly available for testing on November 30, 2022, ChatGPT OpenAI’s chatbot has caused a stir on the internet after breaking records by gaining over 1 million users in just one week of its launch (Ruby, 2023).

ChatGPT has become popular in various industries, including finance, healthcare, and e-commerce, as a way to incorporate advanced natural language processing capabilities into their products and services. Concomitantly, ChatGPT has been gaining traction in Higher Education.

One of the critical features of ChatGPT is its ability to generate human-like text based on a given prompt. This feature is particularly useful for content creation, language translation, and customer service automation tasks. The model has also been trained on a massive dataset of internet text, allowing it to understand and respond to a wide range of topics and use cases. AI-generated text tools like ChatGPT have been trained to use billions of written works, encompassing different materials, including blogs, novels, and classic literature (Marr, 2022).

Regarding language support, ChatGPT can understand and generate text in several languages, including English, Spanish, French, German, and Chinese. However, the model’s performance may vary depending on the language and the complexity of the task requested.

**Potential & Impact**

**ChatGPT in Higher Education**

With its ability to generate human-like text and carry out natural language processing tasks, ChatGPT can potentially impact how students and teachers work in Higher Education. One of the key areas where ChatGPT can make a significant impact is in automating administrative tasks. For example, ChatGPT can answer frequently asked questions and provide course information freeing up valuable time for teachers and allowing them to focus on what they do best: teaching.

ChatGPT has the potential to be highly beneficial in personalized learning support (Trust et al., 2023). By facilitating text-based conversations between students and ChatGPT, learners can interact with an intelligent system that responds in real time, leading to a more personalized and engaging learning experience. ChatGPT can be used in various ways, such as providing instant feedback to students, answering their questions, or generating relevant resources. ChatGPT can act as a “teachable agent,” according to Tate et al. (2023), by engaging students in a text-based conversation and encouraging them to teach ChatGPT a concept as a means of promoting their own learning.

It is also possible to use ChatGPT to promote deep learning by improving student engagement. One example is personalized tutoring, where ChatGPT can provide customized feedback and guidance to individual students based on their learning needs and preferences. This can help students better understand and apply the course content. Another example is scenario-based learning, where ChatGPT can simulate real-life situations and provide students with opportunities to apply their knowledge and skills in an applied context.

In addition to improving student engagement, ChatGPT can help to streamline the grading process, providing real-time feedback and offering feedback on students’ writing (Trust, 2023), allowing students to receive quick and constructive feedback on their progress. This can be especially valuable for online learning, where traditional methods of offering feedback and grading can be slow and cumbersome.

ChatGPT has been used to generate automated summaries and transcripts of online lectures. This allows students to quickly and easily review the key takeaways from an online lecture and provides a valuable resource for students who may have difficulty understanding the spoken word. ChatGPT can also be integrated into online discussion boards or virtual reality environments to facilitate more meaningful interactions between students and instructors. One may argue that students can benefit, particularly in terms of productivity and efficiency. By using ChatGPT, students can streamline their research and study processes, saving time and increasing productivity.

Another relevant aspect of ChatGPT’s potential in Higher Education is language learning. Due to its ability to carry out natural language processing tasks, ChatGPT can facilitate conversational practice, providing students with personalized feedback on their pronunciation and language skills. When integrated into online courses, ChatGPT has the ability to understand and generate text in multiple languages. This allows educators to create multilingual online course materials, making their courses more accessible to global audiences. Additionally, ChatGPT can offer writing and communication assistance for students, particularly those who face language and learning challenges, as well as individuals who are learning English as a second language (Tate et al., 2023).

Looking broadly, we acknowledge the potential of generative AI and integrate it into education, just as we use other tools such as computers, calculators, and pen and paper. By incorporating AI into the curriculum and teaching students how to use it responsibly, we can help them gain a better understanding of its ethical implications in society, formal and informal education, and lifelong learning (Bozkurt et al., 2023).

**Obstacles & Constraints**

**ChatGPT in Higher Education**

Simply put, ChatGPT is an artificial intelligence language model that generates text by analyzing input data. However, it does not have the capability to comprehend the intricacies of human language and conversation fully. Its responses are solely based on the information fed to it during the training process (Bogost, 2022).

An important aspect to consider is that ChatGPT can only offer information dated before September 2021. If there are inquiries regarding topics from before the Internet era, such as a paper on the 1968 flu pandemic, there will be limited sources to draw from. Therefore, it is unsurprising that numerous limitations and constraints of ChatGPT have been pointed out. For starters, when using ChatGPT in Higher Education, one of the obstacles is the lack of context awareness. Since the model is trained on a massive dataset of internet text, it does not have the ability to understand the specific context of the prompt. As a result, it may generate text that is not entirely relevant or accurate and lead to errors or inaccuracies in its generated text. On the same token, ChatGPT can produce content that imitates the style and format of real news stories but incorporates fabricated statistics and quotes from fictitious sources (Vincent, 2019).

Another obstacle faced when using ChatGPT is its lack of creativity. ChatGPT generates text based on patterns it has learned from the data it was trained on, which means it can produce repetitive or generic text. Also, it needs help understanding the tone of a given text, which can lead to inappropriate responses or feedback. This is a constraint when using ChatGPT to generate feedback on student assignments, as it may not be able to understand the nuances of human communication and can generate feedback that is unsuitable to the student’s needs.

Creativity comes into play when the learner or user prompts ChatGPT. That leads to Prompt Engineering or the art of asking ChatGPT questions (Mok, 2023). Prompt engineering is a technique used to help artificial intelligence understand what to say or write. It involves selecting and designing the right words to guide the model toward producing the desired type of response. This technique helps to improve the accuracy and consistency of language models when generating natural language text. Educators are currently working on collecting and sharing ideas on using generative AI tools (e.g., ChatGPT, Dall-E, and Midjourney) to enhance their teaching methods and improve academic research. They collaborate to gather and exchange innovative ideas on incorporating these tools into their scholarship (Nerantzi, 2023).

A significant constraint of generative AI is algorithm bias (Bozkurt et al., 2023). Algorithm bias refers to the tendency of an AI algorithm to produce biased outputs or results that are unfair or discriminatory toward specific groups of people. This can happen if the algorithm is trained using biased data or instructions, leading to biased decisions or predictions. In the case of generative AI, algorithm bias can severely limit its potential. Generative AI systems are designed to create original content, such as images or text, based on the data they have been trained on. For example, if a language model is trained on biased data, it may generate text that perpetuates harmful stereotypes or discriminates against certain populations. Consequently, algorithm bias is a major concern because it can reinforce and amplify existing social inequalities and discrimination (Bembeneck et al., 2021).

**Concluding Thoughts**

**ChatGPT in Higher Education**

With the release of GPT-4 this week by OpenAI, ChatGPT and other forms of generative AI in Higher Education are here to stay. GPT-4 is now able to understand and respond to both images and text, making it a multimodal AI tool. OpenAI reports that the updated version has improved accuracy and expanded knowledge, along with the ability to “see” images and “reason” (OpenAI, 2023).

Generative AI has the potential to revolutionize the way we learn and teach. By leveraging the power of generative AI, educators can provide more personalized and satisfying learning experiences. Students may be able to maximize their productivity and achieve their academic goals in a more efficient manner.

Yet, it is essential to acknowledge the pitfalls of this technology. Algorithmic bias, content quality, currency, and accessibility issues must be addressed to ensure that ChatGPT is used to its fullest without exacerbating existing inequalities. At the end of the day, while technology can significantly enhance the educational experience, it is the human-to-human interactions that make education genuinely transformative.

References

Bembeneck, E., Nissan, R., & Obermeyer, Z. (October 21, 2021). To stop algorithmic bias, we first have to define it. Brookings.

Bogost, I. (December 7, 2022). ChatGPT Is Dumber Than You Think. The Atlantic.

Bozkurt, A. et al. (2023). Speculative Futures on ChatGPT and Generative Artificial Intelligence (AI): A Collective Reflection from the Educational Landscape. Asian Journal of Distance Education, 18(1), 53-130. https://doi.org/10.5281/zenodo.7636568

OpenAI, Brockman, G., & Sutskever, I. (December 11, 2015). Introducing OpenAI. OpenAI Blog.

OpenAI (March 14, 2023). GPT-4 is OpenAI’s most advanced system, producing safer and more useful responses. OpenAI Blog.

Marr, B. (December 21, 2022). ChatGPT: Everything You Really Need To Know (In Simple Terms). Forbes.

Mok, A. (March 1, 2023). ‘Prompt engineering’ is one of the hottest jobs in generative AI. Here's how it works. Insider.

Nerantzi, C. (February 2, 2023). Creating a collection of 101 creative ideas to use AI in education. #creativeHE.

Ruby, D. (February 8, 2023). ChatGPT Statistics for 2023: Comprehensive Facts and Data. DemandSage.

Tate, T. P., Doroudi, S., Ritchie, D., Xu, Y., & Uci, M. W. (2023, January 10). Educational research and AI-generated writing: Confronting the coming tsunami. EdArXiv.

Trust, T. (January 2023). ChatGPT & Education.

Trust, T., Whalen, J., & Mouza, C. (2023). Editorial: ChatGPT: Challenges, Opportunities, and Implications for Teacher Education. Contemporary Issues in Technology and Teacher Education.

Vincent, J. (2019). AI researchers debate the ethics of sharing potentially harmful programs. The Verge.

-- Please cite the content of this blog as: Correia, A.-P. (2023, March 15). Is ChatGPT the new buzz in Higher Education? Ana-Paula Correia’s Blog. https://www.ana-paulacorreia.com/anapaula-correias-blog/2023/3/15/is-chatgpt-the-new-buzz-in-higher-education