Publicado em April 12, 2023

Chat GPT in Education: potential and challenges


In this article, we will discuss some of the potentials and challenges of GPT Chat in Education. In recent years, artificial intelligence (AI) has proven to be a valuable structure, sometimes as a tool and sometimes as a resource, in several fields, including education. One of the ways in which AI has been applied in education is through virtual assistants such as ChatGPT. ChatGPT is a large natural language processing (NLP) neural network, trained to answer questions and provide information on a variety of topics. GPT is one of the most advanced language models available today and is widely used in natural language processing applications such as chatbots, virtual assistants and content recommendation systems.

GPT is an acronym that means "Generative Pre-trained Transformer" in Portuguese has been translated as "automatic generation of redundant text". It is a language model developed by OpenAI, which uses a neural network architecture known as "transformers" to generate natural language text.GPT is pre-trained on large amounts of textual data to learn the relationship between words, and then can be fine-tuned for specific tasks such as answering questions or generating text summaries.

The use of GPT Chat can be used in several areas of knowledge, but its application in education has drawn attention and questions, especially in the context of knowledge production. The use of natural language processing technologies has been increasingly used in education, especially with the rise of distance learning and e-learning platforms. One such technology is GPT Chat, a language model trained to converse with users in natural language.

Chat GPT and its potential in education

One of the key benefits of ChatGPT in education is its ability to provide a personalized learning experience for each student. ChatGPT can adapt to each student's level of knowledge and skills, providing specific information and answers to each question asked.

For now, it is possible to point out that the GPT Chat can be used in several areas and domains of knowledge, and can cover theoretical, practical and cognitive computing skills.

GPT Chat can be used in education in several ways. One of them is the creation of educational chatbots, which may be able to talk to students to clarify doubts, givefeedbacks and provide relevant information. Companies that provide educational services using technology-based business models claim that these chatbots can be used at different levels of education, from elementary school to higher education, and in different areas of knowledge.

The challenges involving GPT Chat in Education

As far as challenges go, it is important to note that ChatGPT cannot completely replace human interaction in education. While ChatGPT can provide contextualized and personalized information, it cannot provide the same level of feedback and guidance that a human teacher can. Additionally, ChatGPT cannot answer questions involving creativity, critical thinking, emotions and values, areas where human interaction is crucial.

GPT Chat can generate many opportunities, but at the same time many challenges in educational contexts. The model can be a useful resource to interact with students and support them in the development of research and study activities. On the other hand, many educators have already begun to show concerns in points that concern the lines of learning, theoretical perspectives, the critical development of the consultant, the theoretical approach and the pedagogical perspective.

So, these are extremely relevant points that need to be taken into account in learning contexts. It is also worth noting that the issues related to plagiarism and copyright are worrying. In this regard, it is necessary to point out that there are still no other anti-plagiarism systems capable of identifying the productions elaborated by the GPT Chat, not to mention that their productions can also be considered plagiarism, since their model when producing a text, on a certain topic, is getting the human knowledge available in cyberspace by collective intelligence to generate a summary.

An article published in Jornal da USP on February 13, entitled: “ChatGPT: the robot that shows how artificial intelligence can revolutionize our lives”, by Osvaldo N. Oliveira Jr., Thiago A. S. Pardo, and Maria das Graças, calls attention to a very interesting fact demonstrating problems in text interpretation. However, the article considers that this is a problem that will soon be solved as the interaction provides more and more learning for the system.

The article also exposes that the technology is a great advance in the context of AI. Oliveira, Prado and Graças (2023) point out that the abilities of the GPT Chat “indicate that we are taking great steps towards the new paradigm for knowledge generation”, where the machines themselves can autonomously generate knowledge about the world and about themselves ( WITHOUT HUMAN INTERVENTION).

It's hard to know how teachers of generations still in the workforce will come to terms with technology like this without applying scientific research. But we can have an idea based on recent technologies, such as cell phones. In this sense, many schools and some teachers had taken the decision to ban the use of cell phones in the classroom. Currently, this has become impossible. I usually say that it is not possible to deny technology. We need to appropriate it and make the best use of it, not the other way around.

Regarding the identification of academic works elaborated by the GPT chat, OpenAI itself points out that it is elaborating an anti-plagiarism system that will be able to identify productions elaborated by the GPT chat. But for now we don't have an anti-plagiarism tool of this level.

279 reads 60 Likes

About the author

Barbara Coelho
I'm too connected!

Barbara Coelho

Research Coordinator, LTI Digital-UFBA

Doctor in Education (UFBA). Post-Doctor in Information Science (UNB). Professor at UFBA. Permanent Professor at PPGE-UFBA. Researcher at PPGCI-UFSCar. Lead researcher at the Informational Technologies and Sociodigital Inclusion Laboratory (LTI Digital - ltidigital.ufba.br). Open Knowledge Foundation Civic Innovation Ambassador.
Atualmente realiza investigação das Aplicações de IA na Educação.

Read too

Quantum Machine Learning (QML)

Quantum Machine Learning (QML)

Among the many peculiarities, the notion of quantum superposition challenges our imagination significantly

279 reads 60 Likes
Café com Presidente
Próximo Evento
June 29, 2023

Café com Presidente

Uma conversa com o Presidente da I2AI - Onédio S. Seabra Júnior - para falarmos sobre os temas mais quentes de Transformação Digital e Inteligência Artificial, num bate-papo informal com