Abstract

This academic research paper explores the possible implications of Artificial Intelligence on modern educational systems. It uncovers how these technologies can revolutionise the learning process, significantly benefiting students and contemporary society. The study argues that the co-existence of human educators and AI-powered tools in settings where the latter acts in the supporting role is the environment that will maximise the potential for improvement of the educational field when considering the latest advancements in the generative AI field. The paper assesses examples of current usage of Artificial Intelligence in the educational space, focusing on their positive effects while considering how existing side effects can be negated. Furthermore, the scholarly analysis establishes a comprehensive framework for educators and researchers in the field of generative AI. The framework should enable them to effectively utilise AI's capabilities and broaden the benefits that AI-powered technologies can bring in the educational context.

Introduction

Over the past decade, Artificial Intelligence has undergone a remarkable surge in growth and advancement. Among the numerous technological breakthroughs within the last seven years, notable creations include Tesla's Autopilot in 2014, Amazon's Alexa assistant in 2014, and Google's AlphaGo playing AI in 2016. Despite these impressive innovations, only the release of ChatGPT in November 2022 has managed to capture the attention of the general society and dispel any doubts about the great potential that AI has.

The extensive adoption of this all-purpose text-generative chatbot, seemingly possessing comprehensive knowledge about our society and capable of generating solutions to any question, removed the veil of scepticism surrounding AI's capabilities and left a remarkable impact on people’s perceptions of computer automation (Chan & Tsi, 2023). It prompted individuals to wonder how the integration of similar tools into current industries might change various career fields where humans constitute the main workforce.

The first collaboration between AI and the education field dates back to the 1950s when the first computer-assisted instructions were created (Nwana, 1990). Despite such a rich history, AI's potential role as a substitute for human teachers has only recently become a topic of increasing interest and discussion.

Thanks to the widespread availability of the internet and continuing industrialisation, at present generative AI tools are much more accessible to the public than ever before (Gillani, Eynon, Chiabaut, & Finkel, 2023). In the modern age, accessing powerful public large-language models has become remarkably easy for anyone. With platforms like ChatGPT for GPT-3.5 and Bing for GPT-4, individuals can readily input a short prompt and receive a generated answer. Despite the ease of usage, academic institutions are generally not embracing the potential of these innovations for educational purposes. Instead, they tend to prevent students from using these tools in their academic journey and are reluctant to integrate such technology into the educational context in any capacity (Schiff, 2021). The studies show that 43% of American undergraduate and graduate students have experience utilising generative LLMs such as ChatGPT, indicating that the learners are keeping up with the changes in society, while academia at large isn’t employing the technology and potentially may lose its ability to prepare students for the real world (Welding, 2023). Academia is struggling to embrace new technology based on Artificial Intelligence, with many educators hesitating to take the first steps toward integrating it into their curriculum. A lot of teachers’ reluctance is explained by their disapproval of AI technology in any form, as they claim that it “can produce original work that sounds more human” and will lead to them fully changing their course structure and the set of skills that they want to teach to the learners (Yorio, 2023). Thus, it causes resentment towards AI technology and opposition to normalising the usage of AI in the educational setting (Yorio, 2023).

The given study reasons that the integration of Artificial Intelligence into the education system has the potential to enhance teaching methods, alleviate educators’ workload, and support learners in their studies. As such, it is a necessary technology that must be utilised in the educational content. The paper argues that the beneficial and effective education of the future is grounded in the harmony between human teachers and AI-powered tools, while the replacement of human teachers with AI is not a threat due to some complex and incomprehensible to AI natures of educators’ duties that require the “human factor”.

The forthcoming sections of the scholarly analysis delve into three key aspects: the consensus in the research field regarding the replacement versus cooperation between human educators and Artificial Intelligence, how AI-powered tools can substantially support teachers in the educational process, and the ethical concerns arising from the potential integration of generative tools. Building upon the insights gathered from these sections, the paper proposes a comprehensive set of suggestions that form a robust framework aimed at making Artificial Intelligence integration possible, without compromising the levels of equality, privacy, and instructing proficiency that are in place in modern teaching institutions nowadays.

Literature on AI and Teachers

A sudden increase in AI task performance capabilities has sparked a large interest in the topic of comparing AI and human educators in the research community. The following section investigates the existing research literature that compares both parties and evaluates the pro et contra of each in the context of education.

Advantages of AI

Some proponents for the integration of AI in classrooms suggest that AI could effectively supplement or even replace human teachers in most contexts. Endorsers of complete AI automation of education argue that AI-powered systems can complete tasks without being affected by human favouritisms or biases. AI's ability to work tirelessly, providing support and personalised guidance to learners without fatigue or time constraints, makes the replacement attractive from a reliability and, point of view (Kasneci, et al., 2023).

Extensive research has demonstrated that AI can automate virtually all administrative tasks within educational institutions without suffering in decreased efficiency or accuracy (Yan, et al., 2023). Apart from that, it is suggested that AI is capable of providing objective feedback to students on their quizzes and assignments. AI's use of algorithms enables it to assess student performance based on predefined criteria, eliminating potential biases and subjectivity often associated with human grading (Tamkin & Ganguli, 2021).

This feature allows for precise evaluation of students' strengths and weaknesses, facilitating timely interventions and tailored learning paths. Throughout the learning year, AI can continuously evaluate student progress and know then the instructional strategies should be adapted accordingly (Wang, Li, Tan, Yang, & Lei, 2023). By analysing vast amounts of data and performance patterns, AI can provide real-time insights into individual and group achievements, thus fostering a data-driven approach to education.

Advantages of Teachers

However, despite its potential benefits, AI's limitations must also be acknowledged. One critical aspect in which AI falls short is self-awareness (Schiff, 2021). Unlike human teachers who possess emotional intelligence and consciousness, AI cannot communicate with humans by effectively using correct emotions, preventing it from forming a genuine understanding of students' emotional needs. Human educators often create a sense of engagement with their students through emotional connections, which enhances the learning experience for the learners (Yorio, 2023). AI, lacking this emotional aspect, may struggle to replicate the empathetic and supportive environment that human teachers can create.

Additionally, there are concerns about the potential adverse effects of reliance on AI as educators on students' social development. It has been researched that AI doesn’t understand human society due to complications in the representation of social norms and insufficient capabilities to quantify the social concepts of humans (Chan & Tsi, 2023). Human interactions play a vital role in shaping social skills, and excessive dependence on AI-driven instruction may hinder students' ability to navigate real-life social contexts, and since the students gain a lot of their social behaviours from the early days of the classroom, it is a detriment to the whole society and especially to the next generation.

General Consensus

All in all, the literature on AI and teachers demonstrates that while AI offers numerous advantages, it is not without limitations. Despite these limitations being perceived by some as minor, it is generally accepted that the educational place must not only pass on the knowledge to the next generations but also uphold certain social values, which is impossible at the current stage of AI development. Thus, it is generally considered that AI is not able to replace human beings in educational roles. Nonetheless, most papers argue that AI can assist human teachers and make learning more effective. As educational institutions explore the integration of AI in classrooms, it is crucial to find a careful balance that maximises the benefits of AI while acknowledging and addressing its limitations. The concept that the current research field is moving towards emphasises the symbiotic relationship between AI and human educators. In this model, AI aids teachers in handling mundane and repetitive tasks, using its capabilities in generalisation (Chan & Tsi, 2023). Meanwhile, human teachers assume a more personalised and emotionally supportive role for the learners (Chan & Tsi, 2023). This complementary approach seeks to maximise the strengths of both AI and educators, fostering a more effective and enriching learning experience.

Applications of AI

After recognising the established strengths of both AI and current teachers, it becomes evident that Artificial Intelligence will not replace human educators. However, to demonstrate the essential role of Artificial Intelligence integration into the education system, it is required to explore the areas where automation can effectively support human teachers and to examine which roles the educators will concentrate on. This investigation will shed light on how AI can be used to enhance and complement the work of educators.

One of AI's most significant advantages lies in its capacity to automate tasks that involve generalisation and repetition. These tasks encompass a wide range of categories, including Profiling/Labelling, Detection, Grading, Prediction, Knowledge Representation, Feedback, Content Generation, and Recommendation, as well as Teaching Support (Yan, et al., 2023). By automating these tasks, AI enables educators to focus on more nuanced aspects of teaching, such as fostering critical thinking, creativity, and personalised student support (Chan & Tsi, 2023).

Case Study: Content Generation

Generative AI's capability in producing content based on a prompt allows it to create personalised educational content tailored to special learning styles and needs, helping educators alleviate the hardships of adapting existing curricula to a specific case. A great example of such usage is David Blumenkrantz's creation of a curriculum for the photojournalism programme of the Kenyan university with the assistance of generative AI, which helped him to produce work that "would have taken months without the AI-generated research" in just a few weeks (Walker, 2023).

Case Study: Labelling

Moreover, AI's proficiency in labelling and detection tasks plays a pivotal role in streamlining the information flow for human teachers, which is utilised by numerous college professors (Walker, 2023). By efficiently sorting and organising vast amounts of data that teachers have to deal with, AI facilitates educators' workflow, empowering them to focus on delivering high-quality instruction and individualised support to their students (Yan, et al., 2023).

Case Study: Teaching Support

As defined by the study of AI task completion capabilities in the educational context, the Teaching Support category consists of types of tasks that are related to “classroom teaching, learning community support, online learning conversation agent, intelligent question-answering, teacher activity recognition” (Yan, et al., 2023). An illustrative example of AI's impact in the educational context as a teaching supporting tool can be observed at the University of Auckland, where in 2021 OpenAI's Codex was adapted to assist students with understanding complex coding errors. This AI-powered tool demonstrated great results, as 88% of students reported finding the feature beneficial in enhancing their understanding and problem-solving skills (Leinonen, et al., 2023).

By utilising AI-powered tools to handle a substantial portion of student enquiries, teachers can significantly reduce their workload. This, in turn, creates more opportunities for educators to concentrate on assisting learners in comprehending intricate concepts. Moreover, such integration of AI will enhance the value of teachers' pedagogical skills, as they can dedicate their expertise and attention to fostering deeper understanding and critical thinking among students. The success of the study that was conducted before generative chatbot LLMs became as capable as they are nowadays demonstrates the big potential of using Artificial Intelligence for teaching support.

Emotional and Ethical Limitations

However, AI does have limitations, particularly in areas that necessitate Emotional and Interpersonal, Pedagogical, and Creativity skills. AI's inability to empathise or engage emotionally with students poses a challenge in cultivating meaningful teacher-student relationships, which play a vital role in enhancing the learning experience (Chan & Tsi, 2023).

Additionally, AI may fall short when it comes to addressing educational topics that require an understanding of Ethics, Morals, and Civics (Yan, et al., 2023). AI currently lacks the capacity for nuanced decision-making, whereas discussions that involve moral dilemmas or complex ethical considerations often demand human reasoning and judgement.

Nonetheless, AI serves as a valuable resource for human teachers. By providing them with vast amounts of data and insights, AI enhances educators' ability to understand their students' learning patterns, preferences, and progress. Armed with this knowledge, teachers can deliver more targeted and personalised instruction, catering to each student's unique needs and learning pace.

All in all, AI's applications in education have proven transformative in automating repetitive tasks and streamlining information management. While AI is not a substitute for the interpersonal and creative skills of human educators, it empowers teachers with valuable data and support, enabling them to optimise their teaching practices and create a more engaging and personalised learning environment for students. As AI continues to advance, it is crucial to forge a harmonious balance between AI-powered technologies and human educators, capitalising on the strengths of each to unlock the full potential of education in the digital era.

Ethical Concerns

The integration of Artificial Intelligence and Large Language Models into the educational field raises a set of significant ethical concerns. This subsection addresses the main ethical considerations related to the use of AI in education, with a focus on Privacy, Equality, and Beneficence.

Privacy Concerns

A primary ethical concern revolves around privacy issues stemming from the use of LLMs trained on student data. It is imperative to ensure that student data is collected with proper consent and adherence to privacy regulations. However, in the current reality, most LLMs that are trained on student data that wasn’t given with the proper consent from students, instead utilising the “opt-in” strategy by having students and their guardians sign an agreement when signing up for some educational service that gathers data in the background (Yan, et al., 2023). Thus, most people don’t even realise that their data is used when training AI-powered tools, raising valid concerns about data protection and confidentiality. Additionally, most LLMs lack sufficient forgetting methods, which could inadvertently lead to these models generating outputs that disclose sensitive and private information (Tamkin & Ganguli, 2021). To address this concern, AI researchers and engineers need to adopt a more human-centric development approach, implementing robust privacy protocols and data management practices.

Equality Concerns

The dominance of the English language in the Natural Language Processing pipeline presents an ethical challenge in terms of access to better generative models. English-speaking individuals benefit from more sophisticated and refined LLMs compared to speakers of other languages (Wang, Li, Tan, Yang, & Lei, 2023). This language disparity may inadvertently widen the gap between Western and Third World Countries, impeding progress towards global equality (Gillani, Eynon, Chiabaut, & Finkel, 2023). To address this concern, efforts should be made to diversify and expand the linguistic capabilities of AI-powered technologies, ensuring equitable access to high-quality generative models for all language communities.

Bias Concerns

Another significant ethical concern pertains to potential bias in LLMs, which may arise due to cost considerations. While developing LLMs can be costly, some institutions may resort to utilising cheaper versions of AI models that have been trained on less controlled or diverse datasets. Although cheaper models do not necessarily imply inferior performance, the concern lies in the possibility of bias creeping into the AI's decision-making processes. This bias may inadvertently favour certain students over others, hindering the learning progress of some individuals and perpetuating educational inequalities (Gillani, Eynon, Chiabaut, & Finkel, 2023). Thus, developers and educators need to conduct thorough evaluations and implement measures to mitigate bias in AI systems, ensuring fair and equitable treatment for all students.

Overall, the ethical concerns surrounding the integration of AI and LLMs in education are significant and require careful consideration by educators. Addressing these concerns necessitates a collaborative effort involving AI researchers and educators. By prioritising data privacy, promoting language diversity, and actively mitigating biases, the educational community can harness the full potential of AI technologies while upholding ethical principles and safeguarding the interests and rights of students in an increasingly AI-driven educational landscape.

Framework

Even though it seems like the Artificial Intelligence industry is starting to shift from research-oriented goals to the creation of products for specific consumer needs, it is evident that there’s a huge gap in communication between machine engineers and educators (Aggarwal, 2021). Despite benefits that could be achieved with machine intelligence in the educational field, it is clear that modern Artificial Intelligence won’t be widely integrated into academic institutions until the technology meets the safety requirements, ethical concerns, and employment status that are placed forward by the human educators. The proposed framework aims to address all of these points without compromising on the benefits that AI-powered technology can bring.

Synergy

The Artificial Intelligence role in the context of education isn’t meant as a replacement for human educators, but rather as a supporter that is meant to alleviate teachers from doing mundane work. This can be represented by grading short answer questions, doing simple research when preparing for a lesson, or taking over administrative tasks.

AI is capable of guiding learners to better understand the course material by answering simple questions, thus making teacher office hours reserved when students have more complex enquiries that require a professional’s explanation. Moreover, AI-powered tools can gather data on how well students are learning in the course, providing teachers invaluable information and helping students to catch up with an understanding of the course before it’s too late.

In this setting, the role of human teachers undergoes a transformation that places greater emphasis on emotional support for students. They take on the responsibility of engaging students with the course content more effectively, assisting when students face challenges in their learning journey, and teaching crucial critical-thinking skills. These areas are where AI currently lacks, making the involvement of human educators vital for creating a well-formed learning experience (Perkins, 2023).

Fairness

Bias in Language Model Models is a significant concern, challenging to control due to various factors such as the lack of data, insufficiently complex neural models, and limited training time (Kasneci, et al., 2023). While many methods exist to manage bias by imposing restrictions on generated content, addressing unfairness and discrimination in the output often requires more training time, additional data, and larger models. However, these enhancements typically demand increased monetary funds.

Despite the limited financing faced by academic institutions, predictions indicate that the cost of training AI-powered tools will continue to decrease significantly in the future (Taulli, 2023). This suggests that, over time, affordable Artificial Intelligence technology for education, with lower bias and appropriate output for educational contexts, will likely become more accessible to academic institutions.

Dialogue

Engaging in open conversations between machine learning engineers and educators holds great significance. These discussions enable a comprehensive understanding of the essential features needed to enhance Artificial Intelligence tools for educational contexts. In particular, the integration of AI-powered tools calls for LLMs that differ from the standard market models like OpenAI’s GPT-3.5 or GPT-4.

One of the primary purposes of these tools is to assist students, which makes models that produce research papers or original work on request useless within academic institutions. Instead, the focus should be on developing Artificial Intelligence that offers hints to students when they enquire and leads them to come to an answer on their own, promoting a more challenging but rewarding learning experience.

Recognising that the AI industry may not possess the same level of familiarity with educational needs as educators themselves, it becomes crucial to engage in conversations that define how AI should function in these contexts. These exchanges of ideas and insights will ensure that AI development aligns effectively with the specific requirements of education, ultimately enhancing students' learning abilities and overall educational experience.

Ethics Education

As the wide adoption of Artificial Intelligence in society continues to grow, academic institutions will inevitably face the need to revamp their curriculum, eliminating assignments that can be easily completed by generative AI models, with writing assignments being particularly susceptible at present. However, amidst these changes, it remains essential to prioritise educating learners about the ethical usage of AI-powered technology in both academic and professional settings.

The creation of modern LLMs has introduced a new wave of academic integrity violations (Perkins, 2023), almost implying that cheating was non-existent in the past. However, it is important to acknowledge that academic cheating has been a concern historically and will continue to be in the future. The root of this issue is not grounded in the existence of AI but lies in the ill intentions of some students.

Academic institutions should persist in their efforts to educate learners about the far-reaching effects of academic cheating on society's future and students' workplace performance. Moreover, they must teach students the crucial difference between ethical and unethical usage of AI technologies. This will enable students to understand the margin better, empowering them to prevent accidental violations of academic rules and foster a culture of responsible AI usage.

Conclusion

In conclusion, the integration of Artificial Intelligence into the education system holds great promise for transforming the learning landscape and improving educational outcomes. While AI has undergone significant growth and advancement in recent years, the consensus in the research field suggests that AI should not be viewed as a replacement but as a complementary tool that enhances the role of human educators.

The applications of AI in education, such as content generation, labelling, and teaching support, have demonstrated its capacity to automate repetitive tasks and streamline information management. By automating these tasks, AI can free up educators to focus on more nuanced aspects of teaching, such as fostering critical thinking and creativity, and engaging the students. The success of the presented case studies in this research highlights the positive impact AI can have on the learning experience.

However, ethical concerns surrounding AI integration also exist. Privacy issues related to data collection and usage, language disparities in AI capabilities, and the potential for bias in AI systems demand careful consideration and proactive measures. Addressing these ethical concerns requires collaboration between AI researchers and educators to ensure the responsible and equitable use of AI technologies in education.

The proposed framework emphasises the importance of synergy between AI and human educators, recognising that AI is not meant to replace teachers but to support them. It advocates for the development of AI tools that promote learning, critical thinking, and problem-solving rather than generating answers upon request. The framework also calls for fairness in AI applications and urges open dialogue between AI engineers and educators to cater AI technologies to the specific needs of education. Furthermore, ethics education becomes a crucial component in preparing students to navigate the academic honesty challenges posed by AI. By educating students about responsible AI usage and academic integrity, institutions can foster a culture of ethical AI adoption and responsible technology usage.

By using AI to augment the teaching process, educators can create a more engaging and personalised learning experience for students. As the AI field continues to evolve, ongoing collaboration, ethical considerations, and responsible AI usage will pave the way for a brighter future in education, where technology and human expertise work hand in hand to empower learners for success in the digital age.

References

• Aggarwal, G. (2021, July 20). Council post: Why are there so few direct-to-consumer AI startups? (Forbes Magazine). Retrieved August 4, 2023, from https://www.forbes.com/councils/forbestechcouncil/2021/07/20/why-are-there-so-few-direct-to-consumer-ai-startups/

• Chan, C. K., & Tsi, L. H. (2023). The AI revolution in education: Will AI replace or assist teachers in higher education?

• Cruz-Benito, J. (2022). AI in Education. MDPI - Multidisciplinary Digital Publishing Institute.

• Gillani, N., Eynon, R., Chiabaut, C., & Finkel, K. (2023, January). Unpacking the “Black Box” of AI in Education. Educational Technology & Society, 26(1), 99-111.

• Hwang, G.-J., Tang, K.-Y., & Tu, Y.-F. (2022). How artificial intelligence (AI) supports nursing education: Profiling the roles, applications, and trends of AI in nursing education research (1993-2020). Interactive Learning Environments, 1-20.

• Kasneci, E., Seßler, K., Küchemann, S., Bannert, M., Dementieva, D., Fischer, F., . . . Kasneci, G. (2023, April). ChatGPT for good? On opportunities and challenges of large language models for education. Learning and Individual Differences, 103. doi:10.1016/j.lindif.2023.102274

• Leinonen, J., Hellas, A., Sarsa, S., Reeves, B., Denny, P., Prather, J., & Becker, B. A. (2023, March 15-18). Using large language models to enhance programming error messages. Proceedings of the 54th ACM Technical Symposium on Computer Science Education, 1. doi:10.1145/3545945.3569770

• Nwana, H. S. (1990, December). Intelligent tutoring systems: An overview. Artificial Intelligence Review, 4, 251-277.

• Perkins, M. (2023). Academic integrity considerations of AI large language models in the post-pandemic era: ChatGPT and beyond. Journal of University Teaching and Learning Practice, 20(2). doi:10.53761/1.20.02.07

• Schiff, D. (2021, September 2). Education for AI, not AI for education: The role of education and ethics in national AI policy strategies. International Journal of Artificial Intelligence in Education, 32(3), 527-563. doi:10.1007/s40593-021-00270-2

• Tamkin, A., & Ganguli, D. (2021, February 5). How Large Language Models Will Transform Science, Society, and AI. (Stanford University). Retrieved July 31, 2023, from https://hai.stanford.edu/news/how-large-language-models-will-transform-science-society-and-ai

• Taulli, T. (2023). Generative AI: How ChatGPT and other AI tools will revolutionize business. Berkeley: Apress.

• Walker, R. (2023, June 14). California professors test out AI in the classroom, even as cheating debate continues. (CalMatters). Retrieved August 4, 2023, from https://calmatters.org/education/higher-education/2023/06/ai-classroom-california/

• Wang, X., Li, L., Tan, S. C., Yang, L., & Lei, J. (2023, September). Preparing for AI-enhanced education: Conceptualizing and empirically examining teachers’ AI readiness. Computers in Human Behavior, 146. doi:10.1016/j.chb.2023.107798

• Welding, L. (2023, March 27). Half of College Students Say Using AI on Schoolwork Is Cheating or Plagiarism. (BestColleges). Retrieved July 4, 2023, from https://www.bestcolleges.com/research/college-students-ai-tools-survey/

• Xie, H., Hwang, G.-J., & Wong, T.-L. (2021). Editorial Note: From conventional AI to modern AI in education: Re-examining AI and analytic techniques for teaching and learning. Educational Technology & Society, 24(3), 85.

• Yan, L., Sha, L., Zhao, L., Li, Y., Martinez-Maldonado, R., Chen, G., . . . Gašević, D. (2023). Practical and ethical challenges of large language models in education: A systematic scoping review.

• Yang, S. J. (2021). Guest editorial: Precision education - A new challenge for AI in education. Educational Technology & Society, 24(1), 105-108.

• Yorio, K. (2023). The ChatGPT revolution: School librarians explore new AI technology. Will it dramatically change education? School Library Journal, 69(2), 10.