В этой статье автор проводит исследовательский анализ тревожной тенденции использования искусственного интеллекта в процессе изучения языков. В связи с таким уникальным, но вполне ожидаемым развитием событий в этой области возникает вопрос о том, не вытесняют ли генеративные модели искусственного интеллекта преподавателей. Автор ищет ответы на этот вопрос, анализируя работы других исследователей на схожую тему.

Ключевые слова: ИИ, генеративный искусственный интеллект, преподаватель.

Since time immemorial teaching has been a crucial part of everyday life of a person whether they be the one who provides the education or the one who receives it. Pedagogy in its core presents a variety of challenges that surround the core of knowledge residing within, nevertheless learners all over the world and time have attempted to create shortcuts or easier ways to obtain knowledge or circumvent the educational process. Such a process is quite ironic when viewed through the lens of language teaching, since as such these people resemble language itself and its nature that strives to be as efficient as possible with time it uses for communication. With that idea in mind, it is obvious to see how some people may view AI and subsequently the tools powered by it when it comes to teaching. This is especially poignant when looking at language teaching specifically. When given a task, students may use Artificial Intelligence tools to create vast swathes of text for their written assignments as well as avoiding the learning process in a variety of other ways. While the specifics of AI use regulation makes for a topic of its own, the more important aspect of this problem, which is also the focal point of the present paper, is the fact that not only is this a blatant way for learners of all ages to avoid the educational load but also a big problem for pedagogical specialists, especially language teachers. The two questions that come as a result of such a way of think is quite simple: if AI can generate grammatically sophisticated texts, translate, explain vocabulary and produce written texts, what aspect of language learning should remain central and what learning means, when a machine can produce the language for the learner. In order to answer these puzzling and seemingly simple but incredibly complex examples of a linguistic and philosophical conundrum, the present article will focus on analysis of works of several other researchers who wrote on a similar or specifically on the mentioned subject.

As it has already been stated, the main and the focal point of the problem many teachers may face when confronted with AI tools and their subsequent “fruits of labor” is actually the fact that there is nothing that a pedagogical expert can do that has not been by AI interference. Even though this does sound quite simple and primitive at first the problem, as it turns out, is so basic in its nature only due to the fact that the main concept of teaching as a whole is affected. It is no surprise that with the increase in AI usage various researchers have stated their own theories and opinions on its fair use and probable necessity. As a prime example is the work of Felice M. and her colleagues that can be seen in the article made for the British Council in which they ponder the question of fair AI use and ethics surrounding the subject. Their main ideas circle around the potential that artificial intelligence tool may bring to various fields of work and expertise. The researchers note in their work several times that automated tools such as ChatGPT can be misused or be connected to immoral actions, and that such a thing is not happening seldomly. However, with that said, it is also stated that the whole use or misuse issue of the AI tools can be fixed with the introduction of ethics and that the implementation of artificial intelligence instruments into almost every faucet of our lives is inevitable. As such instead of policing the eventual and inevitable future that is directly intertwined with the use of these quite often despised instruments, the authors suggest moral and ethical use under fair guidelines. The reason for this empathetic and embracing outlook is pushed due to the fact that AI tools, while nevertheless able to be used in malicious or destructive ways, can prove to be an invaluable ally to a person striving to consume knowledge or to better themselves.

These statements are all the more serviceable as arguments in the main question of the present work. While it is the absolute objective truth that AI can be used by learners of any type to cut corners or simple avoid doing tasks either due to their inability to accomplish a task or due to their simple laziness, if implemented with moral and ethical guidelines, the tools can instead be used to empower the participants of the learning process to do more with their educational time. This, however, requires all of the members of the pedagogical activities, both from the side of learners and those who provide the education, to be aware of the AI use at the very least, as well as be prepared to give the other party more leeway with their methods and showing trust.

Besides trust it is also important to note that regulation plays in big role at a much bigger scale to secure the aspects of GenAI and its tools. This is why the work of Holmes W. for UNESCO on human-centered approach for GenAI in regard to regulation creation is another good article to analyze. While the author focuses not only on their main point and also goes on to define GenAI as a concept, what should catch anyone’s eye is the list of factual problems stemming from the use of artificial intelligence use in the field of education. Among others Holmes notes worsening digital poverty, use of content without consent, AI-generated content pollution, lack of understanding the real world, reduction in diversity of options and further marginalizing already marginalized voices and generating deeper deepfakes. All of these problems, with some being more obscure than others at first glance due to the unique naming of the issues, focus mainly on the same thing when it comes to learning and education: those who may attempt to use AI tools in the educational processes must be aware of dangers that such a path poses, meaning one may come upon untrustworthy writings sourced from unknown origins, or even worse, plagiarized unbeknownst to both parties involved. While these issues are described by Holmes as complete doom and gloom, in actuality according to the author with the implementation of a variety of regulations with human-centric focus on the governmental level, the educational field may actually benefit from use of AI as well.

The article provides several examples of this that include methods on how to include GenAi in the teaching process as a helper symbolically, as it can be used to design curriculums and courses with increased efficiency or as a teaching assistant. All the while it can additionally be used by people to serve as a personal coach for the self-paced acquisition of foundational skills in learning languages and the arts and those who require special needs in their educational process. Once again, one may find a similar meaning of the general understanding of AI. It is a double-edged sword that can be used positively to indulge people’s strive for knowledge or negatively as a tool that propagates laziness and cheating in its truest form.

The ethical implementation of AI in the educational field is quite similar to what back in the day less tech-enabled participants had to endure. While notetaking, cheat sheets and other more basic ways of circumventing the traditional educational process never left the scene, AI nevertheless takes center stage as a ringleader since even if learners do not use these artificial intelligence tools to necessarily cheat, they deprive both themselves and their teaching specialists charged with providing them education of valuable experience that would otherwise be available to them.

When analyzing AI tools and their potential in a more positive light, one can easily come to a conclusion that with the use of generative content one can fulfill any task or create any basic text that is required of the person, however that is simply untrue. When taking a look language-wise generative artificial models may actually not only give themselves away by their speech or writing style, which is actually good in the context of teaching, but also may lack a specific flair that may be required of the student to complete an assignment. Just like mankind’s progress the language evolution never sits still either and with each passing year or even sometimes month or week, the language receives alterations or new vocabulary or grammatical rule or shortcut to keep up with the necessary communicational demand that in of itself is stemming from the rapid development people as a global society see and live in every day. As such even with how some of the adaptive generative AI models work, communication can still have something up its sleeve that cannot be replicated fully or at least partially by artificial intelligence. This is exactly what Lee S. writes about in their article “Generative AI and English language teaching: A global Englishes perspective”. In this article the author analyses how AI views English forms and how its linguistic integration of English may come up challenges that machines are yet to overcome.

English as a language is used worldwide by people from enormously diverse linguistic and cultural backgrounds. While English language basics are set in stone, English language itself isn’t solely used by native speakers and is a global communication tools, which can be even in some capacity be called a language of globalized world. This creates a precedent in which generative models like ChatGPT base their speech patterns on the more traditional and basic English based on mainline American and British version but forgoes the more cultural or context-dependent aspects of this multifaceted language. The authors argue that GenAI creates a potentially interesting problem here because AI models tend to reproduce standardized forms of English and can default towards what the article describes as native-English-speaker norms. For their research the authors of the article have compared the speech, or more specifically writing, patterns of three ChatGPT models with each being more and more refined as they introduced additional specific task-oriented details into their prompts. As per results of their experiments, the conclusion was that while AI model itself strived for more generic native-speakers terms even when given complex detailed prompts, it could still be used to create unique, original and generally outstanding from the English norm texts but that would be heavily dependent on the linguistic assumptions and prowess, as well as data and pedagogical framework that affected prompt. As such it does not actually produce something out of the norm unless specifically worked on by the user and as such requires direct input that may vary in complexity and without proper knowledge or know-how it cannot be relied upon by those who learn the language.

While it may still be argued especially with the results of the latter article in mind that AI tools do make teachers obsolete in their primary designation, the reality is actually far from this statement. While it is true, that generative artificial intelligence can be used at a certain capacity to learn a language, such a task does not actually teach a person the language. As a teacher, one’s responsibility lies not only in making sure the learner knows correct basic forms but also when and how to use them. Moreover, it is not only what one says but how they say it that makes a person a speaker. The culture of an interlocutor and a speaker, their hierarchical status, small country and national varieties in English, contextually defined meaning, language diversity — all these things are critical for a person’s identity as a speaker. The bigger question here is not whether the person can make AI produce a text, but more so can it produce a text that reflects the person and their style, their own, as it has been called before in this paper, flair. With the results provided by Lee and their colleagues, although generally speaking the answer to this question is yes, this in no way makes the teacher’s position obsolete as they have never been there to exclusively peddle grammar and lexical norms. While those are the basics on which any language speaker may proverbially stand, the true identity of the speaker no matter what language resides in how they can and choose to speak in each and every situation that may differ contextually or by any other aforementioned reason. Teachers additionally to basic knowledge of a language ideally contribute heavily in the learner’s understanding of what makes the language tick and how to use it appropriately. This means that things such as language choice evaluation, recognition of different varieties of English, comprehension of contextual meaning, and even in the more advanced classes specifics of English used by multilingual speakers are all aspects of the language that only a teacher can give to their student. This also does not mean in any way that a teacher cannot use AI in their classes either as a tool for preparation or their helper during classes, since with the expert input and acquired knowledge teachers can use AI to circumvent more human-related issues of teaching to produce quality content in shorter time spans.

To sum up, while the more functional or “grunt work” in language studies can be easily replicated by a variety of generative artificial intelligence models the identity of the speaker and their ability to interact with a plethora of external factors affecting their linguistic and communicational choices is something that can only be taught by an educated specialist that provides the necessary tools for further development. And as such there should be no doubt that a person’s position as teacher is not going to be made obsolete by artificial intelligence any time soon.

References:

1. Felice M. et al. Human-centred AI: lessons for English learning and assessment. — 2025.
2. Holmes W. et al. Guidance for generative AI in education and research. — Unesco Publishing, 2023.
3. Lee S. et al. Generative AI and English language teaching: A global Englishes perspective //Annual Review of Applied Linguistics. — 2025. — С. 1–24.
4. Crompton H. et al. AI and English language teaching: Affordances and challenges //British journal of educational technology. — 2024. — Т. 55. — №. 6. — С. 2503–2529.
5. Edmett A. et al. Artificial intelligence and English language teaching: Preparing for the future. — 2023.