Introduction

Artificial intelligence (hereinafter, "AI") systems offer necessary online learning and teaching aid, including customized learning for students, reducing the teacher’s mundane tasks, and formulating adaptive assessments. However, while the positives for AI seem overwhelming, the impact of AI systems on the norm and expectations about interactions between students and teachers is still elusive. In online learning, learner–teacher interaction (i.e., communication, support, and presence) profoundly impacts students' fulfilment and learning results. Thus, recognizing how students and teachers perceive the impact of AI systems is essential to acknowledge any rifts, challenges, or obstacles preventing AI systems from achieving their proposed potential. Although AI systems have been positively credited for improving communication, providing timely and personalized support for wider settings, and enhancing the feeling of connection, there were anxieties about its responsibility, agency, and surveillance issues.

Newer Versions of AI in Education

The opportunities for artificial intelligence (AI) in online learning and teaching are broad, ranging from personalized learning for students and automation of teachers’ routine tasks to AI-powered assessments. For example, AI tutoring systems can provide personalized guidance, support, or feedback by tailoring learning content based on student-specific learning patterns or knowledge levels. AI Teaching Assistants help teachers save time answering students’ simple, repetitive questions in online discussion forums, and instead, teachers can dedicate their saved time to higher-value work like conducting tailored discussions and more insightful dialogue on concerned topics, covering a wider selection of topics. AI analytics allows teachers to understand students’ performance, progress, and potential by decrypting their data. There has been a tremendous development of AI’s concerned with the improvement of online education mechanisms. For instance, a recent AI named Jill Watson has been formulated to augment the teacher’s communication with students by autonomously responding to student introductions, posting weekly announcements, and answering routine, frequently asked questions. Other newer developments seek to build an AI scoring system that allows faster communication of grades between students and the teachers and support them by providing constant feedback on how students learn and the progress they are making towards their learning goals. 

Contemporary Opinions and Concerns

There has been a tremendous increase in research concerning the impact of AI systems in online education. For example, academicians Roll and Wylie call for more involvement of AI systems in the communication between students and teachers, and in education applications outside school context. At the same time, Zawacki-Richter and other experts conducted a detailed systematic review of AI-Education publications from 2007 to 2018. It was found that a lack of critical reflection of the ethical impact and risks of AI systems on learner–teacher interaction and uncovered potential conflicts between them, such as privacy concerns, changes in power structures, and excessive control. All of these studies called for more research into the impact of AI systems on learner–teacher interaction, which will help us target any challenges, or dilemmas hampering AI systems from accomplishing their intended potential.

Indeed, learner–teacher interaction plays a crucial role in online learning. Researchers Kang and Im demonstrated that factors of learner–teacher interaction, such as communication, support, and presence, improve students’ satisfaction and learning outcomes. The learner–teacher interaction further affects students’ self-esteem, motivation to learn, and confidence in facing new challenges. It is important to note that there is minimal information regarding how introducing AI systems in online learning will affect learner–teacher interaction. However, it can be deduced that AI systems would have a deep impact in the classroom, changing the relationship between teacher and student. Although, there is a need to further understand how and why various forms of AI systems affect learner–teacher interaction in online learning.

While the opportunities for AI seem largely positive, students and teachers may perceive the impact of AI systems negatively. For instance, students may perceive indiscriminate collection and analysis of their data through AI systems as a privacy breach, as illustrated by the Facebook–Cambridge Analytica data scandal. Moreover, the behaviour of AI agents that do not take into account the risk of data bias or algorithmic bias can be perceived by students as discriminatory. Teachers worry that relying too much on AI systems might compromise the student’s ability to learn independently, solve problems creatively, and think critically. Therefore, it becomes crucial to determine how students and teachers perceive the impact of AI systems in online learning environments.

Weighing the Pros and Cons        

Firstly, communication refers to questions and answers between students and the teacher about topics directly related to learning contents, such as instructional materials, assignments, discussions, and exams. Students and teachers expect AI systems will positively impact the quantity and quality of communication between them but bears the risk of causing miscommunication and responsibility issues. Students hold that the anonymity afforded by AI would make them less self-conscious and, as a result, allow them to ask more questions. Although students believe AI systems would enhance instructional communication, they worry that AI could give unreliable answers and negatively impact their grades. Despite the fact that AI systems improve instructional communication, students were concerned about responsibility issues that could arise when AI’s unreliable and unexplained answers lead to negative consequences. For instance, when communicating with an AI Teaching Assistant, the black-box nature of the AI system leaves no choices for students to check whether the answers from AI are right or wrong. 

Secondly, support refers to the teacher’s instructional management for students, such as providing feedback, explanations, or recommendations directly related to what is being taught. Students believe that AI would support personalized learning experiences, particularly with studying and group projects. Though AI systems enable tailored support, there is a risk of over-standardizing the learning process. Despite the fact that students appreciate the support they could potentially receive from AI systems, students also worry that standardized support would have a negative influence on their agency over their own learning. 

Thirdly, even though AI strengthens the perceived connection between students and teachers, students are uncomfortable with the measurement of their unconscious behaviour, such as facial expression analysis or eye-tracking, as it feels like surveillance. Therefore, establishing clear, simple, and transparent data norms and agreements about the nature of data being collected from students is an important consideration for future research. To summarise, recent use of AI systems in online learning showed that careless application can cause surveillance and privacy issues, which makes students feel uncomfortable. Therefore, students perceive the impact of AI systems as double-edged swords. 

Conclusion

Various tech communities have promoted claims that digital education enhances access, learning, and collaboration. The COVID-19 pandemic tested these claims like never before, as higher education systems seemingly overnight had to move teaching online. The scholarly literature hints at the complexity involved in implementing educational AI technology for learners, teachers, and institutions. At first glance, several of the positive claims made by AI advocates resonated with student experiences yet numerous questionable attributes are highlighted on deeper analysis. These contradictions in the promises and realities of AI education must be addressed through collaboration that goes beyond a pedagogical focus, enabling transformational change of education mechanisms. 

This article aims to delineate the theoretical implications for a learner–teacher interaction framework by highlighting and mapping key challenges in AI-related ethical issues. It further increases our understanding of the boundaries that determine the student’s trust and acceptance of AI systems, and provide a theoretical background for designing AI systems that positively support bias-free interactions. Although several ethical frameworks and professional codes of conduct have been developed to mitigate the potential dangers and risks of AI in education, significant debates continue about their specific impact on students. In order to minimize the negative impact of AI systems on learner–teacher interaction, it is important to address tensions where AI systems violate the boundaries between students and teachers i.e., responsibility, agency, and surveillance issues. Hence, currently, AI systems are the divider as well as the bridge-builder simultaneously, however, future holistic developments have the potential to highlight the latter rather than the former.

The views expressed above are solely of the author's.