Unscrambling the Contention of Artificial Intelligence and Education

With an imaginary story in mind, a 10th-grade student walked into his writing class, with a mix of hope and apprehension in his heart. The source of his unease? Today, he was about to receive his narrative writing assignment, and the anticipation was almost unbearable. As he crossed the doorway into the classroom, his eyes locked the teacher’s eyes, only to be met with a stern, unwavering expression. His mind raced with concern, and his steps felt heavy as he placed his backpack and moved towards his desk. Just then, the unmistakable ring of the bell marked the beginning of class, sealing his fate. The teacher rose to her feet and commenced the distribution of the writing assignments. For each student, there was a unique blend of devastation and joy as their papers found their way into their trembling hands. Finally, the 10th grade boy received his paper.

What he saw crushed him: a disheartening zero out of 100. Looking up from his paper, his eyes met the teacher’s eyes again. That same glare deepened. Shifting his attention back to his paper, he noticed a blaring red comment from the teacher that declared, ‘You received a zero because you used an AI to write this assignment for you.’ He had been caught, and in that instant, a whirlwind of questions surged through his mind: “Why is using AI to write this assignment harmful?”

Much like this perplexed student’s question, researchers and educators have been embroiled in ongoing debates concerning the use of artificial intelligence in education. Should artificial intelligence be used in education or not? This paper aims to dissect the intricacies, advantages, and challenges associated with the integration of artificial intelligence in education, with a specific focus on Intelligent Tutoring Systems (ITS). To comprehend this topic thoroughly, it is essential to dispel misconceptions and establish a foundational understanding of artificial intelligence. Moreover, gaining insight into the ‘two-sigma problem’ will contextualize the purposed solution of an ITS. With this background, we will explore the advantages and challenges that confront students and teachers.

As we embark on this exploration, it’s crucial to grasp the true essence of artificial intelligence and dispel common misconceptions. In popular media, artificial intelligence often appears as a malevolent super-robot bent on taking over the world, fostering a negative perception that deviates from the genuine nature of artificial intelligence. This distorted view is perpetuated by various news outlets, such as Science News, which features articles like ‘How ChatGPT and Similar AI will Disrupt Education’ (Hulick, 2023). While the concerns raised by Science News may hold validity, this critical portrayal of artificial intelligence predisposes us to bias before we research the advantages and consequences of its implications.

Dissecting this misconception, artificial intelligence is defined as the process in which a machine takes in, interprets, and makes predictions based on data that is given to it. Based upon those predictions, the artificial intelligence refines its algorithm to achieve the results that the artificial intelligence engineer desires.

Pavlo Antonenko and Brian Abramowitz further define artificial intelligence as a machine learning algorithm that seeks out patterns within a large data set, citing examples like “image processing, Google’s route finding, and pattern recognition software” (Antonenko & Abramowitz, 2023). This definition of artificial intelligence exemplifies the process of taking in the data, processing the data, and finding a pattern within the data provided.

With this understanding of artificial intelligence, it is essential to recognize its inherent limitations. Given that machine learning and other artificial intelligence models, such as deep neural networks, demand substantial amounts of specific data for training, their functionality becomes limited and specialized, restricting them to single-use cases that can’t be transferred elsewhere. These requirements and specializations make it challenging to implement artificial intelligence systems, leading developers to create multiple systems for a single problem rather than having one system handle it all.

Although artificial intelligence may have some limitations, the educational system is flawed as well. Within Edward L. Vockell’s article, he summarizes a problem that educational psychologists Benjamin Bloom discovered. While studying student’s learning abilities, Benjamin Bloom noticed that students with individualized tutoring would, on average, score two standard deviations higher on tests than students with conventional instruction methods. As he gathered more data and other researchers concluded similar results, Benjamin Bloom classified this problem as the “two-sigma problem” (Vockell, 1994). Furthermore, the shortage of teachers in the United States amplifies this problem. With schools fighting for teachers and students’ potentials being minimized, it is crucial to find a solution to this issue.

Currently, a purposed solution to this “two-sigma problem” that researchers are exploring is an artificial intelligent tutoring system. This system, as Pablo Arnau-González et all explain, would provide a personalized tutoring experience for students. It does this by replicating human emotions, interactions, and pedagogical approaches to instruct students within a particular subject (Arnau-González et al., 2023). This personalization of learning is what Benjamin Bloom said would increase students learning by two standard deviations. Coupling this potential of artificial intelligent tutoring systems and the traditional teaching of teachers, students would outperform the average America citizen.

With the impact of intelligent tutoring systems, what is holding researchers and school administrators back from implementing it? Are there challenges that come with it? Researchers and school administrators have many benefits and challenges to weigh when it comes to the decision of implementing intelligent tutoring systems. The first set of advantages and challenges originates from students’ use of an intelligent tutoring system.

In an academic article published by Nabeel Gillani et al, students will see benefits from intelligent tutoring systems in three areas of education: subject material learning, test and homework feedback, and academic guidance counseling. First, Nabeel et all state,” ITS … seek[s] to adapt to students’ existing knowledge and skills, or learning states, to help them build skills in more personalized ways.” (Gillani et al., 2023) With an individual intelligent tutor for each student, the tutor adapts to the best learning style for the student and builds their knowledge from where they are at. This is an advantage when part of the class is ahead or on track in the subject, while a few students may be lagging in learning. Having individualized intelligent tutors, the lagging students would have their needs met, while allowing the other students to learn at their pace.

Secondly, Nabeel et al researched the effects that an artificial intelligence system could have on test and homework feedback. Stated in the article, a system could provide instant feedback to students’ homework, deepening their learning upon subject material that they missed. Mainly, this type of feedback system is used with writing assignments, allowing students to understand stronger sentence structure, grammar mistakes, and spelling errors throughout their work. This assessment of work gives students the opportunity to correct and enhance their writing skills (Gillani et al., 2023). Although these systems aren’t explicitly stated as intelligent tutoring systems in the article, the one-on-one experience with artificial intelligence is, in effect, the same pedagogical approach to learning. Additionally, these systems have the potential to help students to creatively think about writing topics that interest them, thus deepening their commitment and desire to write.

Like the assessment artificial intelligence system, the aspect of individualized academic guidance from artificial intelligence mimics an intelligent tutoring system. As Nabeel et al explains, artificial intelligent systems have the potential to guide students down an academic path that captures their interests, strengths, and goals in life. Using an artificial intelligent chatbot, students can ask questions about universities and careers. Additionally, they can find the applications needed for a university (Gillani et al., 2023). By giving high quality insight, this efficient and inexpensive approach to guidance counseling allows students to have less anxiety about careers and higher-level education.

The advantages of intelligent tutoring systems are evident, but students will also face challenges with these systems. From Lauraine Langreo article, challenges that students will face while using artificial intelligence include false information and dependency upon artificial intelligence. In her first problem of false information, Lauraine explains how machine learning artificial intelligence is often trained on out-of-date information. Like the recent Chat-GPT being trained with data from the internet before 2021, intelligent tutoring systems would only be effective up to the date in which they were trained on. Additionally, Lauraine talks about these systems generating incorrect information (Langreo, 2023).

With a system that can generate incorrect and out-of-date information, students’ learning is affected. They won’t be learning accurate information causing misconceptions. Additionally, when new information is found through research, an artificial intelligence system will lack the ability to be updated quickly. These two things will ultimately contribute to students learning information that will negatively impact their learning in the future.

In Lauraine’s second problem, some students will look at artificial intelligence as a tool that does the work for them. Having this view of artificial intelligence, students will create a dependency upon artificial intelligence instead of learning the material themselves. It is important that students learn to master artificial intelligence and use it as a tool rather than the solution (Langreo, 2023). This is important if we want artificial intelligence to create students that outperform traditional learning by two-sigma. If students depend upon artificial intelligence to do their work, then the method a school is using to implement artificial intelligence needs to be redesigned.

Adding upon Lauaine’s problems, other problems that artificial intelligence and intelligent tutoring systems purpose is the lack of human connection. This lack of human connection and the psychological effects are unknown. In the long term, students may be more depressed or express further problems. Unfortunately, research for this challenge is non-existent. It is imperative that researchers take this factor into account for implementing artificial intelligence into education.

With students witnessing most of the advantages and challenges of intelligent tutoring systems, teachers are the second set of people to see the advantages and challenges of intelligent tutoring systems. Beginning with the advantages, in an article written by Kausalai Wijekumar et al, a few advantages include teaching feedback and critical one-on-one student interactions. Building upon the concept of feedback systems discussed for students, teachers can also leverage intelligent tutoring systems to receive feedback on their teaching methods. Kausalai refers to this as “practiced-base- professional development” (Wijekumar et al., 2022). This type of system helps teachers to improve the quality of their teaching and help teachers to look for better ways to engage students in learning.

In addition to feedback systems, Kasualai et al talk about better one-on-one student interactions. Since most of the one-on-one teaching is done by an intelligent tutoring system, teachers can focus on students that are struggling more than usual. This allows teachers to directly impact the lives of students in a one-on-one scenario (Wijekumar et al., 2022). Whether that student has a problem at home, in the classroom, or within the subject, the teacher is still the most important part of the classroom. They have the power to help students get back on their feet or be crushed in defeat. In this critical one-on-one situation, the teacher will see the beauty of teaching. That beauty being the change in students’ lives.

These advantages are great; however, teachers pose more challenges than advantages. When it comes to intelligent tutoring systems, the understanding of technology and teachers’ willingness to implement intelligent tutoring systems are the main problems that researchers and educators face. In a composition of research papers, Helen Crompton et al explain that educators lack of understanding may result in worthless implementation of intelligent tutoring systems or other artificial intelligence systems. In addition, the ability in which teachers can use technology will hinder the performance of intelligent tutoring systems (Crompton et al., 2022).

With teachers and educators having a lack of understanding, attempts to implement intelligent tutoring systems or other artificial intelligence systems will be scuffed. Additionally, students will not receive the potential benefits that may come with intelligent tutoring systems. These systems’ requirements of technological knowledge may cause artificial intelligence systems to be pushed off and ignored, until an educator has the skills to implement such a system.

Furthermore, in an article written by Shanyong Wans et al, they explain that teachers have a negative perception to implementing artificial intelligence within education (Wang et al., 2020). Additionally, Helen Crompton et al found similar results to the teacher and administration willingness to implement intelligent tutoring systems and even go as far to say it is a “self-fulfilling prophecy” (Crompton et al., 2022). For researcher to implement artificial intelligence into education, they need to provide advantages that teachers and administrators can get onboard with. Without the support of faculty, the project is doomed for failure.

All in all, intelligent tutoring systems within K-12 education have many different advantages and challenges. Whether schools will embark on this journey, researchers and educators have a plethora of challenges to solve and advantages to weight. With this ongoing debate, the use of artificial intelligence in education is an enriched discussion that could impact the way we learn. Both the advantages and disadvantages need to be considered, with much thought given before schools move forward with the implementation of artificial intelligence into the educational setting.

References

Antonenko, P., & Abramowitz, B. (2023). In-service teachers' (mis)conceptions of artificial intelligence in K-12 science education. Journal of Research on Technology in Education, 55(1), 64-78. https://doi.org/10.1080/15391523.2022.2119450

Arnau-González, P., Arevalillo-Herráez, M., Luise, R. A., & Arnau, D. (2023). A methodological approach to enable natural language interaction in an intelligent tutoring system. Computer Speech & Language, 81, N.PAG. https://doi.org/10.1016/j.csl.2023.101516

Crompton, H., Jones, M. V., & Burke, D. (2022). Affordances and challenges of artificial intelligence in K-12 education: A systematic review. Journal of Research on Technology in Education, , 1-21. https://doi.org/10.1080/15391523.2022.2121344

Gillani, N., Eynon, R., Chiabaut, C., & Finkel, K. (2023). Unpacking the "black box" of AI in education. Journal of Educational Technology & Society, 26(1), 99-111. https://doi.org/10.30191/ETS.202301_26(1).0008

Hulick, K. (2023, April 12,). How ChatGPT and similar AI will disrupt education. Science News https://www.sciencenews.org/article/chatgpt-ai-artificial-intelligence-education-cheating-accuracy

Langreo, L. (2023). No, AI won't destroy education. but we should be skeptical. Education Week, 43(4), 20-23. https://search.ebscohost.com/login.aspx?direct=true&AuthType=ip&db=asn&AN=171935114&site=ehost-live&scope=site

Vockell, E. L. (1994). The minus two-sigma problem: Correcting defective instruction. Contemporary Education, 65(4) https://www.proquest.com/scholarly-journals/minus-two-sigma-problem-correcting-defective/docview/1291764685/se-2?accountid=4488 http://www.lib.byu.edu/jfinder.pl??url_ver=Z39.88-2004&rft_val_fmt=info

Wang, S., Yu, H., Hu, X., & Li, J. (2020). Participant or spectator? comprehending the willingness of faculty to use intelligent tutoring systems in the artificial intelligence era. British Journal of Educational Technology, 51(5), 1657-1673. https://doi.org/10.1111/bjet.12998

Wijekumar, K. K., Harris, K. R., Graham, S., & Lei, P. (2022). A teacher technology tango shows strong results on 5th graders persuasive writing. Educational Technology Research & Development, 70(4), 1415-1439. https://doi.org/10.1007/s11423-022-10117-9