The terminology of Artificial Intelligence (AI) has been harkened for the past two decades. AI, or learning machine if we may call it so, is believed to be well-distinguished and a cutting-edge computer technology that can assist humans in problem analysis. The technological advancement of AI dumbfounds many contemporary laymen insofar as leveraging its function to be able to do and think like humans do. Russell and Norvig (2010) briefly define and categorize AI into four human-like activities. Firstly, it is a machine with minds that think in a full and literal sense. Secondly, it is a computer that can perceive stimulus, reason, and act. Thirdly, AI is a machine that needs intelligence when it is done by humans. Lastly, it is concerned with intelligent behavior in artifacts. This technology, of how the elements of intelligence advance learning approaches and reasoning to solve problems, has considerably awed ubiquitous internet users – unexceptionally, students and educators.
AI has additionally been used in education, expecting students to experience unique and visual learning that helps them quickly grasp what is being learned. As advancing as it is, China with its 730 million internet users, for example, sets a national AI strategy in its national growth of education (Jing, 2018). Moreover, the Education Ministry allocates 8% of the educational budget at least for supporting the development and application of digitalization in schools or higher education institutions. Although the development of AI is not as rigorous as that in China or Western countries, developing countries such as Indonesia has begun to launch several applications seemingly helpful for improving student learning outcomes. Furthermore, the use of some AI applications started to be prevalent and believed to improve students’ creativity and engagement in learning. Quiver, for instance, is an augmented reality application used to develop picture-coloring activity that seems to come alive and real. Another example is EPIC, a digital reading platform where reading becomes less tedious but more attractive and intriguing with its talking book and quiz games.  Other augmented reality AI tools such as Active Floor and Virtual Reality also take part in bringing learners to experience in-depth learning. For higher educations, Grammarly, Hemingway Editor, and ProWritingAid for instance are commonly used to improve the quality of writing as in grammar-editing, paraphrasing, and vocabulary enrichment to help students develop writing.
Leveraging AI towards improving learning, UNESCO (2019) promotes AI to support the Sustainable Development Goal (SDG) 4, that is to ensure a qualified education for all and support student autonomy in lifelong learning. Intelligent tutoring system and collaborative learning are managed to facilitate all learners, not only students who officially study in schools, but also those who are outside schools – marginalized people, refugees, and people with disability. AI can help educators to reach those who need to learn from home, even hospitals. It opens windows for easy and quick accessible learning. As such, collaborative learning is enhanced with AI to connect those who are not in the same location. In lower education, the benefit of AI has insofar passed the student-teacher trajectory, leading teachers to perceive what is good from AI applications in education. With the help of AI, teachers spend more time working to guide their students with difficulties one-on-one. Considering the considerable amount of time used on administrative tasks, grading tests, assignment preparation, and other extra works, AI frees up some of the teachers’ time to maximize their responsibility in guiding their students. Thus, these experiences of using AI can be considered powerful and, therefore, bring a new paradigm in developing traditional education.
The discussion of the use of AI has been a prolonged debate in the past decade. How beneficial AI may sound, it renders new ways of thinking, opportunities, potentials, and even challenges in current and future education. Consequently, it requires an in-depth understanding of the application of AI in reality for teaching and learning. Teaching and learning should be based on approaches that are suitable for both teachers and learners – especially learners in getting their positive learning outcomes. Thus, in understanding the application of AI more deeply, one is better at attempting to comprehend the three paradigms proposed by Ouyang and Jiao (2021), wherein teaching and learning approaches are associated. These paradigms help and assist AI users to reflect on how far they have used it responsibly.
Ouyang and Jiao (2021) developed the Three Paradigms after an elongated research. Paradigm One talks about AI-Learner relation where AI is directing the learners which is based on the teaching learning approach of Behaviorism. Here, learners merely act as recipients, meaning no cognitive activity is made. All feedback, problem solving, analysis in achieving learning goals are directed from AI tools. The first paradigm notifies that there is one-directed information. Learners only follow the directed path given by AI as the teaching machine. Thereafter, Paradigm Two is getting more into student-centered learning, reclining to the cognitive and social constructivism approach. This approach affords learners an opportunity to reflect that their learning happens when they have mutual interaction with other in situated contexts. Thus, learners herein have cooperated with AI to optimize their learning. AI has no control over learners, instead learners use it as a supporting tool and work with the system during the learning process. Paradigm Three is much more adaptive than the previous two. It takes learners as leader of their own learning, while AI is an additional tool, a small part in a complex and larger adaptive system of learning. It is from a more human perspective who use AI to augment their intelligence by assisting with a higher level of effectiveness. In this context, learners are the ones who decide and solve problems to ensure data accuracy.
What can one reflect from understanding these paradigms? Which perspective is well-thought-of? Would AI be that empowering to learning when, in Paradigm One, learners passively receive notions? There may be a myriad of queries to the use of AI. How to use AI tools should be based on the contextual practicality of each application or AI Tool – specifically in what functions that the tools can offer. While there are AI tools that perform weak specific tasks such as solving arithmetic problems or grading, it is possible that AI would outperform and be able to hold nearly all types of cognitive tasks. In the near future, if human allows AI to control most of the tasks, there is a possibility that AI might bring unbearable risks to humans, specifically to the cognitive development of learners.
ChatGPT
There are plenty of AI instances, as mentioned earlier. Another sample taken from General Artificial Intelligence (GAI) or strong AI is ChatGPT (Chat Generative Pre-trained Transformer). The platform is available online for free and can be accessed by everyone and everywhere easily. Its purpose is to communicate with learners not only to provide information but also to analyze and reason. Its analytical and reasoning ability generates information from reading literature on the internet and responding to questions (Misra & Chandwar, 2023). In an educational framework, some may take the benefits of ChatGPT exploration as being a content generator that makes significant contributions in reporting, emailing, writing stories and poems, and even becoming a virtual language translator.  ChatGPT is considered to be less time-consuming and relinquishing language barriers (Nazir & Wang, 2023). Some recent researches concedes the supporting function of the platform, motivating students to use it during the process of learning and giving good impact in creativity, mathematical and critical thinking, and problem solving. Virtual guidance given to students in difficulties is believed to lead to improved motivation and learning achievement  (Firat, 2023; Yilmaz & Yilmaz, 2023).
Although ChatGPT supports quick feedback or research and provides a larger range of topics at a time, many have been found using it beyond functioning the platform as a supplemental resource, creating the facts that ChatGPT can think and do rationally like humans causes students to fully depend on its capabilities. For some novice students and researchers, merely asking and finding answers from the platform is an allurement. The current idea of writing assignment with the help of dialogic AI is indeed leading some universities to start instructing the students to complete assessments back in tune with traditional handwritten (Misra & Chandwar, 2023). Grounding on the fact that having written text simply taken from the platform is ultimately considered as plagiarism. Other research findings mentioned in Karthikeyan (2023) have strictly warned against the use of ChatGPT platform due to its danger of extreme plagiarism at all levels. Contradictive to the previous studies mentioned in the early paragraph, although as unique as it is, ChatGPT writing product is deemed to be biased for its stringent vocabulary and linguistic choices. Some outputs require further verification since its answers were justified to be mistaken. In case this is all followed passively by students without thinking critically, ChatGPT may have misled students and brought down their creativity to a null. Â Â Â Â Â
In the academic field, writing is one skill normatively learned at all levels and across subjects. Students of all ages, beginning with younger ones, have received writing stimulation to develop cognition – starting to write phonemes, words, sentences, paragraphs, short simple texts, and ultimately larger units of texts where coherence is reflected for others to read and understand. Developing a coherent and effective text is difficult because it is building a whole system of meanings with linguistic components. This requires a high level of creativity and students’ hard work. Thus, composing an extensive text relates to critical thinking to solve problems. Kellogg (2008) emphasizes that when writers retrieve everything stored in the brain, it scaffolds for thinking. Thus, writing and thinking are inseparable twins. Recognizing this literature, many professionals think that the use of ChatGPT is not recommended in the field of education. Other professionals, such as poets, book authors, journalists, lawyers, and accountants are also against the platform (Karthikeyan, 2023). It obstructs the students’ cognitive development, raising fears of degrading human capabilities, engagement, and in the worst case, the loss of self-awareness and conscience of humanity.
