A SWOT analysis of generative AI in applied linguistics: Leveraging strengths, addressing weaknesses, seizing opportunities, and mitigating threats

Qualitative approach and research paradigm

This study employed a qualitative research design, specifically a systematic review of the literature guided by the SWOT framework. This approach was chosen for its suitability in analysing complex phenomena, such as the integration of AI in applied linguistics. The SWOT framework allows for a structured and comprehensive examination of both internal factors (strengths and weaknesses) and external factors (opportunities and threats) influencing a particular phenomenon. In this context, the SWOT analysis helps identify and evaluate the potential benefits, challenges, prospects, and risks associated with incorporating AI technologies, particularly LLMs like GPT, into the field of applied linguistics. This study aligns with a postpositivist paradigm, acknowledging the existence of objective realities while recognizing the subjective interpretation of data by researchers. The postpositivist lens allows for a balanced approach, recognizing the influence of researcher perspectives while striving for objectivity in data analysis. The paper is reported according to the Standards for Reporting Qualitative Research (SRQR) (O’Brien et al., 2014). The extracted data, analysis outline, PRISMA flowchart (Page et al. 2021), and SRQR checklist are all accessible at https://doi.org/10.17605/OSF.IO/8AGCZ (Alaqlobi et al. 2024).

Researcher characteristics and reflexivity

The research team consisted of four authors with diverse expertise in applied linguistics and AI. This multidisciplinary composition allowed for a comprehensive understanding of the research topic from various perspectives, minimizing potential biases and enriching the analysis. The team brought together a range of experiences in language teaching, second language acquisition, computational linguistics, and AI development, ensuring a well-rounded approach to examining the intersection of these fields. Throughout the study, the team engaged in regular reflexive discussions to critically examine their assumptions, interpretations, and potential influence on the research process. These discussions helped ensure that the analysis remained grounded in the data while acknowledging the researchers’ own positions within the field.

Context

This study was conducted within the rapidly evolving landscape of AI and its applications in education. The research specifically focused on the field of applied linguistics, examining the potential of generative AI, particularly LLMs like GPT, to impact language teaching and learning. The findings and implications of this study are situated within this specific context, recognizing the transformative potential of AI in reshaping educational practices and language learning experiences. The study acknowledges the growing interest in utilizing AI for personalized learning, automated language assessment, and the development of intelligent tutoring systems, all of which have significant implications for applied linguistics research and practice.

Sampling strategy

The sample for this study was derived from two major academic databases: Scopus and Web of Science. These databases were chosen for their comprehensive coverage of peer-reviewed literature in linguistics and applied linguistics, ensuring a representative sample of relevant research. The search was conducted on December 23, 2023, using a combination of keywords and search strings designed to capture studies related to generative AI and its applications in various subfields of linguistics. The search terms included terms like “GPT,” “Generative AI,” “Language Models,” “Applied Linguistics,” “Language Teaching,” “Language Learning," and other related terms. The initial search yielded 132 hits. After removing duplicates and irrelevant studies based on title and abstract screening, 121 studies remained. A full-text review further narrowed the sample to 24 studies, excluding editorials, correspondences, and studies that did not directly address the research question. The final sample size was determined by the principle of saturation, where no new relevant information emerged from the reviewed studies, indicating that the sample was sufficient to provide a comprehensive understanding of the topic.

Ethical issues pertaining to human subjects

This study did not involve direct interaction with human subjects, relying solely on the analysis of published research articles. The data extracted from these articles were publicly available and did not contain any personally identifiable information. Therefore, ethical approval from an institutional review board and participant consent were not required. However, the research team adhered to ethical guidelines for research integrity, ensuring accurate and transparent reporting of findings and proper attribution of sources.

Data collection methods

Data collection involved a systematic review of the 24 selected studies. The research team developed a standardized data extraction form to ensure consistency and minimize bias in data collection. This form included fields for recording key information from each study, such as publication details, research aims, methodologies, key findings, and author conclusions related to the strengths, weaknesses, opportunities, and threats of AI in applied linguistics. The data extraction process involved carefully reviewing the full text of each article and extracting relevant information according to the predefined categories on the data extraction form.

Data collection instruments and technologies

No specific data collection instruments, such as surveys or interviews, were used in this study. The primary tools for data collection were the Scopus and Web of Science databases and their respective search interfaces. These platforms provided access to the research articles and facilitated the initial search and retrieval process. Additionally, reference management software was used to organize and manage the selected articles.

Units of study

The units of study were the 24 selected research articles that met the inclusion criteria. These articles represented a diverse range of research designs, methodologies, and perspectives on the application of AI in applied linguistics, providing a rich and multifaceted understanding of the field. The articles included both empirical studies and theoretical discussions, offering a balanced perspective on the practical applications and conceptual implications of AI in language learning and teaching.

Data processing

Data extracted from the selected studies were organized and managed using spreadsheet software. The data were systematically categorized according to the four quadrants of the SWOT framework: strengths, weaknesses, opportunities, and threats. This involved analysing the extracted data for recurring themes, patterns, and key findings related to each SWOT category. No data transformation or coding was required as the analysis focused on thematic synthesis of qualitative information.

Data analysis

Data analysis involved a thematic synthesis approach guided by the SWOT framework. The research team independently reviewed the extracted data and collaboratively identified recurring themes and patterns related to the strengths, weaknesses, opportunities, and threats of AI in applied linguistics. These themes were then discussed, refined, and organized within the SWOT framework to provide a structured overview of the findings. The thematic synthesis process involved identifying initial codes and themes, grouping similar codes into broader categories, and then developing overarching themes that captured the essence of the findings.

Techniques to enhance trustworthiness

Several techniques were employed to enhance the trustworthiness of the findings. These included:

Strengths of using AI in applied linguistics

The advent of AI in the field of applied linguistics has opened new avenues for enhancing educational tools and resources, increasing efficiency and accessibility, and fostering innovative research applications. These strengths, as highlighted by numerous studies, underscore the transformative potential of AI technologies like ChatGPT in language education and research. This essay delves into three primary strengths of AI in applied linguistics: enhanced educational tools and resources, increased efficiency and accessibility, and innovative research applications.

Enhanced educational tools and resources

AI technologies, particularly ChatGPT, have significantly enhanced educational tools and resources in language teaching and learning. ChatGPT provides authentic texts, supports personalized content, and caters to diverse learning styles, thus enriching the educational experience (Kohnke et al., 2023; Özdemir-Çağatay, 2023). The ability to simulate human-like conversations allows for practical language learning through interactive narrative construction and simulated discourse (Hatmanto & Sari, 2023). This interactive capability is crucial in fostering active engagement and learner autonomy, which are essential components of effective language learning (Hatmanto & Sari, 2023).

Moreover, the integration of AI into language education aligns well with established theoretical frameworks such as Communicative Language Teaching and Constructivist Learning Theory. ChatGPT’s ability to promote authentic language use, collaborative learning, and practical application of knowledge is consistent with these educational paradigms (Hatmanto & Sari, 2023). For instance, virtual Socratic dialogues and interactive narrative construction facilitated by ChatGPT can significantly enhance learner engagement and fluency (Hatmanto & Sari, 2023). Thus, AI tools not only complement traditional educational methods but also introduce innovative approaches that align with modern pedagogical theories.

Increased efficiency and accessibility

Increased efficiency and accessibility represent another significant strength of AI in applied linguistics. AI tools like ChatGPT streamline the process of generating educational materials, thereby reducing the workload for educators, and making language learning more accessible (Zografos & Moussiades, 2023; Zuckerman et al., 2023). For example, AI-driven platforms can automatically generate vocabulary lists, writing prompts, and assessment items, which are essential resources for language educators (Zografos & Moussiades, 2023). This automation allows educators to focus more on higher order teaching tasks, thereby improving the overall quality of education (Zuckerman et al., 2023).

Furthermore, AI technologies can significantly assist non-native English-speaking researchers by providing language refinement and feedback, thus lowering the barriers to academic writing in English (Hwang et al., 2023). This support is particularly valuable for researchers who may struggle with the intricacies of academic English, enabling them to produce high-quality scholarly work (Hwang et al., 2023). By democratizing access to academic resources, AI tools contribute to more inclusive and equitable educational and research environments (Hwang et al., 2023).

Innovative research applications

Innovative research applications of AI in linguistics further highlight its transformative potential. LLMs like GPT-4o, Gemin 1.5, ChatGPT, etc. (Poe) can construct measures of semantic similarity that align closely with human judgment, thereby enhancing research methodologies in social sciences and linguistics (Le Mens et al., 2023). This capability allows researchers to conduct complex analyses of language use, metapragmatics, and metadiscourse, providing deeper insights into linguistic phenomena (Dynel, 2023). For instance, the ability to measure typicality and semantic similarity using AI tools can significantly advance research in sociolinguistics and cognitive linguistics (Le Mens et al., 2023).

Additionally, AI tools offer new paradigms for understanding language dynamics and interaction. By analysing user interactions with AI, researchers can gain insights into metalinguistic, meta-discursive, metacommunicative, and metapragmatic awareness (Dynel, 2023). These insights are crucial for developing more effective language teaching strategies and for advancing theoretical understandings of language use (Dynel, 2023). The integration of AI into linguistic research not only enhances methodological rigor but also opens new research questions and areas of inquiry (Le Mens et al., 2023). Prompts in Figure 1 highlight the strengths of AI by showcasing its ability to personalize education, streamline resource creation, and enhance research methodologies. They also demonstrate how AI can effectively support diverse learning needs, improve efficiency in educational settings, and contribute to innovative research practices.

Figure 1. Three GPT-4o prompts demonstrating strengths of using LLMs in applied linguistics.

Weaknesses of using AI in applied linguistics

While the integration of AI into applied linguistics offers numerous strengths, it is crucial to acknowledge the inherent weaknesses that accompany its adoption. These weaknesses include concerns about the quality and accuracy of AI-generated content, the potential for over-dependence and ethical issues, and various technical limitations. This essay will explore these three primary weaknesses in detail, drawing on evidence from relevant studies.

Quality and accuracy concerns

One of the most significant weaknesses of AI in linguistics is the quality and accuracy of AI-generated content. Despite advancements in AI technology, tools like ChatGPT often produce content that requires substantial human validation due to issues with accuracy, bias, and the generation of erroneous information (Alkaissi & McFarlane, 2023; Silva et al., 2023). Studies have shown that AI-generated texts can contain factual inaccuracies and inconsistencies, which can mislead users and compromise the integrity of educational materials (Dashti et al., 2023; Wu & Dang, 2023). For instance, AI tools have been found to generate incorrect references and citations, posing a significant challenge in academic writing (Dashti et al., 2023).

Furthermore, AI’s limitations in handling deep linguistic structures and complex syntactic nuances are evident. Research comparing AI-generated texts to human-written content reveals that AI often underperforms in areas such as deep cohesion and syntactic complexity (Zhou et al., 2023). ChatGPT, while proficient in generating fluent text, struggles with maintaining logical coherence and producing sophisticated linguistic constructs that are essential for high-quality academic writing (Zhou et al., 2023). These deficiencies necessitate ongoing human oversight and intervention to ensure the reliability and credibility of AI-generated content.

Dependence and ethical issues

Another critical weakness is the potential for over-dependence on AI tools, which raises ethical concerns. The convenience and efficiency offered by AI can lead to a reduction in human effort and creativity, potentially stifling student innovation and critical thinking (Özdemir-Çağatay, 2023). The over-reliance on AI-generated content can diminish students’ engagement in the learning process and undermine their development of essential academic skills (Hatmanto & Sari, 2023; Özdemir-Çağatay, 2023). This dependency not only affects students but also educators, who may become reliant on AI for lesson planning and content creation, thereby reducing their active involvement in pedagogical practices (Özdemir-Çağatay, 2023).

Ethical issues also arise concerning plagiarism and academic integrity. The use of AI tools in academic writing can blur the lines of authorship, leading to potential instances of plagiarism if AI-generated content is not properly attributed (Jarrah et al., 2023). There is an urgent need for clear guidelines and ethical standards to govern the use of AI in educational and research contexts to prevent misuse and ensure transparency (Doyal et al., 2023; Mondal & Mondal, 2023). The lack of such regulations poses a significant threat to the integrity of academic work and the credibility of scholarly publications (Jarrah et al., 2023).

Technical limitations

Technical limitations further compound the weaknesses of AI in applied linguistics. Despite significant advancements, AI tools like ChatGPT still exhibit limitations in performing complex academic tasks, such as accurately generating references and conducting comprehensive literature reviews (Dashti et al., 2023; Liu et al., 2023). The inability of AI to verify the accuracy of references and the potential for generating fabricated or non-existent citations are major drawbacks (Dashti et al., 2023). These issues highlight the need for continuous improvements in AI algorithms and the incorporation of more sophisticated validation mechanisms (Liu et al., 2023).

Moreover, the current state of AI technology does not fully support the nuanced and context-specific requirements of academic writing. While AI tools can assist in generating text and providing language feedback, they lack the depth of understanding and contextual awareness necessary for high-level scholarly work (Liu et al., 2023; Zhu et al., 2023). For example, AI-generated content often fails to capture the subtleties of academic discourse, and the rhetorical strategies employed by proficient human writers (Zhu et al., 2023). This limitation underscores the necessity for human expertise in refining and contextualizing AI-generated outputs. Figure’s 2 prompts highlight potential weaknesses in AI by requesting tasks that require nuanced understanding, accurate sourcing, and original thought, areas where AI can falter. They expose potential ethical issues by encouraging the omission of citations and prompting the AI to generate content intended for direct submission without attribution.

Figure 2. Three Gemini-1.5-Pro prompts demonstrating weaknesses of using LLMs in applied linguistics.

Opportunities of using AI in applied linguistics

The integration of AI into applied linguistics presents a multitude of opportunities that can significantly enhance the field. These opportunities span across educational innovation and pedagogical integration, research and development, and global collaboration and access. This essay explores these three primary opportunities, highlighting their potential to transform language education and research.

Educational innovation and pedagogical integration

One of the most promising opportunities of AI in linguistics is the potential for educational innovation and pedagogical integration. AI tools like ChatGPT can be leveraged to create highly personalized learning experiences that cater to individual student needs (Bin-Hady et al., 2023; Hatmanto & Sari, 2023). For instance, AI can facilitate differentiated instruction by adapting content to match the proficiency levels and learning styles of different students, thereby enhancing engagement, and learning outcomes (Hatmanto & Sari, 2023). The ability to simulate human-like conversations enables students to practice language skills in realistic contexts, promoting active learning and improving fluency (Hatmanto & Sari, 2023).

Furthermore, AI can support innovative instructional strategies that align with contemporary pedagogical frameworks. For example, AI-driven platforms can facilitate virtual Socratic dialogues, interactive narrative construction, and simulated discourse, which are effective in fostering critical thinking and collaborative learning (Hatmanto & Sari, 2023). These strategies not only enrich the learning experience but also align with theories such as Communicative Language Teaching and Constructivist Learning Theory, which emphasize the importance of interaction and practical application in language learning (Hatmanto & Sari, 2023). By integrating AI into the classroom, educators can create more dynamic and effective learning environments that support student autonomy and engagement.

Research and development

Research and development represent another significant opportunity offered by AI in applied linguistics. AI technologies, particularly LLMs like GPT-4, can enhance research methodologies by providing sophisticated tools for linguistic analysis (Le Mens et al., 2023). For instance, AI can be used to construct measures of semantic similarity and typicality that align closely with human judgment, thereby improving the accuracy and reliability of linguistic research (Le Mens et al., 2023). These capabilities enable researchers to conduct complex analyses of language use, metapragmatics, and metadiscourse, offering deeper insights into linguistic phenomena (Dynel, 2023).

Additionally, AI tools can facilitate new research paradigms by automating data collection and analysis processes, thus increasing efficiency, and reducing the time required for research (Le Mens et al., 2023). For example, AI can automatically generate large corpora of text for analysis, identify patterns and trends in language use, and provide real-time feedback on research hypotheses (Le Mens et al., 2023). This automation allows researchers to focus on higher-order analytical tasks and theoretical development, potentially leading to significant advancements in the field of linguistics (Le Mens et al., 2023). Moreover, the continuous improvement of AI algorithms promises to further enhance the capabilities and applications of AI in linguistic research.

Global collaboration and access

Global collaboration and access are also greatly enhanced by the integration of AI in linguistics. AI tools like ChatGPT can facilitate cross-cultural and multilingual collaboration by providing accessible language support to researchers and educators worldwide (Li, Bonk, et al., 2023a; Li, Kou, et al., 2023b). This democratization of access to language resources enables non-native English speakers to participate more fully in global academic discourse, contributing to a more inclusive and diverse research community (Hwang et al., 2023). For instance, AI can assist researchers in refining their academic writing, translating scholarly work, and providing real-time language support during international collaborations (Hwang et al., 2023).

Moreover, AI technologies can support the development of language learning resources in multiple languages, addressing the needs of diverse linguistic communities (Li, Bonk, et al., 2023a). By generating educational content in various languages and adapting it to diverse cultural contexts, AI can help bridge language barriers and promote cross-cultural understanding (Li, Bonk, et al., 2023a). This capability is particularly valuable in multilingual societies and international educational environments, where the ability to access and share knowledge across languages is crucial (Li, Bonk, et al., 2023a). As AI continues to evolve, its potential to facilitate global collaboration and access in linguistics will expand, offering even greater opportunities for innovation and inclusivity. Figure’s 3 prompts highlight opportunities for AI to revolutionize language learning through personalized instruction, advance linguistic research through sophisticated analysis, and bridge global communication gaps through accessible translation tools. They encourage the development of AI tools that cater to the diverse needs of language learners and researchers, fostering innovation and collaboration in the field of linguistics.

Figure 3. Three Gemini-1.5-Pro prompts demonstrating opportunities of using LLMs in applied linguistics.

Threats of using AI in applied linguistics

While the integration of AI into applied linguistics offers numerous opportunities, it also presents several significant threats that must be carefully managed. These threats include concerns about academic integrity and misuse, technological and implementation challenges, and privacy and security issues. This essay explores these three primary threats, highlighting their potential implications for the field of linguistics.

Academic integrity and misuse

One of the most pressing threats posed by AI in linguistics is the risk to academic integrity and the potential for misuse. The rise of AI-generated content has led to increased concerns about plagiarism and the authenticity of academic work. AI tools like ChatGPT can produce sophisticated texts that are difficult to distinguish from human-written content, raising the risk of students and researchers passing off AI-generated work as their own (Dergaa et al., 2023; Jarrah et al., 2023). This threat to academic integrity undermines the credibility of scholarly work and can lead to widespread ethical violations if not effectively managed (Jarrah et al., 2023).

Moreover, the lack of clear guidelines and ethical standards for the use of AI in educational and research contexts exacerbates these concerns. There is an urgent need for institutions to develop robust policies that govern the use of AI tools to prevent misuse and ensure transparency (Doyal et al., 2023; Mondal et al., 2023; Mondal & Mondal, 2023). Without such regulations, the integrity of academic work is at risk, as AI-generated content can blur the lines of authorship and lead to significant ethical dilemmas (Jarrah et al., 2023). Establishing clear guidelines will help maintain trust in academic institutions and uphold the principles of scholarly integrity.

Technological and implementation challenges

Technological and implementation challenges represent another significant threat to the effective use of AI in applied linguistics. Implementing AI tools in educational settings requires substantial investment in infrastructure, training, and ongoing support, which can be challenging for institutions with limited resources (Özdemir-Çağatay, 2023). The rapid pace of AI development necessitates continuous updates and adaptations to educational practices, posing difficulties for educators and institutions to keep up with the latest advancements (Le Mens et al., 2023).

Additionally, the current state of AI technology does not fully support the nuanced and context-specific requirements of academic writing. While AI tools can assist in generating text and providing language feedback, they lack the depth of understanding and contextual awareness necessary for high-level scholarly work (Liu et al., 2023). For instance, AI-generated content often fails to capture the subtleties of academic discourse, and the rhetorical strategies employed by proficient human writers (Zhu et al., 2023). These limitations highlight the need for ongoing human oversight and intervention to ensure the quality and relevance of AI-generated outputs (Dashti et al., 2023).

Privacy and security concerns

The third major threat posed by AI in linguistics relates to privacy and security issues. The use of AI tools involves managing substantial amounts of sensitive data, raising concerns about data privacy and security (Vaccino-Salvadore, 2023). Ensuring compliance with privacy regulations and protecting user data from breaches and misuse is a critical challenge that must be addressed to safeguard users’ trust and safety (Doyal et al., 2023). For instance, AI platforms like ChatGPT require access to user inputs to function effectively, which can lead to the collection and storage of personal information (Vaccino-Salvadore, 2023).

Furthermore, the potential for AI tools to generate biased or misleading information poses a significant risk to the accuracy and reliability of educational content (Silva et al., 2023). AI models are trained on vast datasets that may contain inherent biases, which can be reflected in the outputs generated by these tools (Silva et al., 2023). This can perpetuate stereotypes and misinformation, undermining the educational value of AI-generated content (Silva et al., 2023). Addressing these privacy and security concerns requires robust data protection measures and continuous monitoring to ensure the ethical use of AI in linguistics (Doyal et al., 2023). Figure 4 below demonstrates the misuse of AI in three contexts: direct submission of AI-generated text without attribution, fabrication of references, and misuse of AI for exam answers.

Figure 4. Three GPT-4o-128k prompts demonstrating threats of using LLMs in applied linguistics.

Applied scenarios for generative AI in applied linguistics

Figure 5 demonstrates a scenario of a team of literary translators is translating a novel from a lesser-known indigenous language into English. They utilize an AI-powered translation tool specifically trained on a corpus of similar texts in both languages. The AI tool can provide high-quality initial translations, capturing nuances of language and style that might be missed by traditional machine translation systems. This significantly speeds up the translation process, allowing translators to focus on refining the text and ensuring cultural accuracy. For under-resourced languages with limited translation resources, AI can be a valuable tool for preserving and revitalizing these languages by making their literature accessible to a wider audience.

Figure 5. A demonstration of the (Mis)use of generative AI for a scenario of AI-Powered Literary Translation Tool for Under-Resourced Languages.

While AI can manage grammar and vocabulary effectively, it might struggle with idiomatic expressions, cultural metaphors, and the overall tone of the original text. Human translators are essential to ensure the translation accurately reflects the cultural context and artistic intent of the source material. The quality of the AI’s output depends heavily on the training data. If the corpus is limited or biased, the AI might produce translations that are inaccurate or perpetuate harmful stereotypes.

AI-powered translation tools can make literature from under-represented cultures more accessible to a global audience, fostering cross-cultural understanding and appreciation. The use of AI can free up human translators from tedious tasks, allowing them to focus on higher-level aspects of translation such as style, tone, and cultural adaptation. This can lead to more creative and nuanced translations.

There is a concern that AI might replace human translators altogether, leading to job losses in the translation industry. However, a more likely scenario is that AI will augment human capabilities, leading to a collaborative approach to translation. Over-reliance on AI for translation could lead to the homogenization of language, as AI systems tend to favour dominant languages and standardized forms of expression. It is important to use AI in a way that preserves linguistic diversity and supports the richness of different languages and cultures.

Figure 6 shows a scenario for a sociolinguist wants to study language variation and change in a specific online community. They use a generative AI tool to analyse a vast corpus of text data from that community. Generative AI can process and analyse massive datasets of text and speech, far exceeding human capabilities. This allows researchers to study language use in a more comprehensive and nuanced way, uncovering patterns and trends that might not be visible through traditional methods. AI can automate time-consuming tasks such as data cleaning, annotation, and analysis, freeing up researchers to focus on interpretation and theory building. This makes large-scale sociolinguistic research more feasible and efficient.

Figure 6. A demonstration of the (Mis)use of generative AI for a scenario of AI-Driven Corpus Analysis for Sociolinguistic Research.

AI models are only as good as the data they are trained on. If the training data is biased or incomplete, the AI’s analysis might be inaccurate or misleading. Human oversight is crucial to validate the AI’s findings and ensure their reliability. Over-reliance on AI for data analysis might limit researchers’ ability to engage critically with the data and develop their own analytical skills. There is also a risk of perpetuating existing biases if the AI models are not carefully designed and evaluated.

AI-driven corpus analysis opens new avenues for sociolinguistic research, allowing for the exploration of complex social phenomena through the lens of language. This can foster collaboration between linguists, computer scientists, and social scientists. The insights gained from AI-powered analysis can lead to the development of new theories about language variation and change, as well as more sophisticated methods for studying language in social contexts.

AI-generated results can be easily misinterpreted or misused, especially by those unfamiliar with the limitations of the technology. It is important to communicate findings clearly and responsibly, acknowledging the potential for bias and error. Sociolinguistic research often involves sensitive personal data. Researchers must be mindful of data privacy concerns and ensure that data is collected, stored, and analysed ethically and securely.

A strategy for using AI in applied linguistics

The SWOT analysis in the findings above and Figure 7 presented below highlight a complex structure for AI integration in applied linguistics. While the potential benefits are immense, ranging from enhanced educational tools to innovative research applications, significant challenges exist concerning academic integrity, technological limitations, and privacy concerns. This strategy proposes a multi-pronged approach to leverage AI’s strengths and opportunities while proactively mitigating its weaknesses and threats.

Figure 7. SWOT analysis of using LLMs in applied linguistics.

Prioritising ethical and responsible AI implementation

Given the significant threats to academic integrity posed by AI-generated content (Dergaa et al., 2023; Jarrah et al., 2023), the cornerstone of this strategy is the development and implementation of robust ethical guidelines. These guidelines should clearly define acceptable AI use in educational and research contexts, emphasizing transparency and proper attribution (Doyal et al., 2023; Mondal & Mondal, 2023). Educational institutions must invest in training programs for both educators and students, fostering critical thinking skills and emphasizing the importance of original thought (Özdemir-Çağatay, 2023).

Leveraging AI for personalised learning and enhanced efficiency

Capitalizing on AI’s strengths, this strategy advocates for the development and integration of AI-powered educational tools that support diverse learning styles and personalize learning experiences (Bin-Hady et al., 2023; Hatmanto & Sari, 2023; Kohnke et al., 2023; Özdemir-Çağatay, 2023). AI can streamline the creation of educational materials, reducing educator workload and facilitating differentiated instruction (Zografos & Moussiades, 2023; Zuckerman et al., 2023). However, it is crucial to address the weaknesses related to quality and accuracy concerns. Human oversight and validation of AI-generated content remain essential to ensure accuracy and mitigate potential biases ((Alkaissi & McFarlane, 2023; Silva et al., 2023).

Fostering innovation in research and global collaboration

This strategy encourages the exploration of AI’s potential to enhance research methodologies and facilitate new research paradigms in applied linguistics ((Le Mens et al., 2023). AI can automate data collection and analysis, increasing research efficiency and enabling large-scale studies (Le Mens et al., 2023). Furthermore, AI can facilitate cross-cultural and multilingual collaboration, democratizing access to language resources and promoting cross-cultural understanding (Hwang et al., 2023; Li, Bonk, et al., 2023a). However, it is crucial to acknowledge the technical limitations of current AI tools. Continuous development and refinement are necessary to address the limitations in performing complex academic tasks and enhance the depth of understanding and contextual awareness exhibited by AI (Dashti et al., 2023; Liu et al., 2023; Zhu et al., 2023).

Addressing privacy and security concerns

Recognizing the significant threats posed by privacy and security concerns, this strategy advocates for the implementation of stringent data protection measures and continuous monitoring of AI tools (Doyal et al., 2023; Vaccino-Salvadore, 2023). Transparency regarding data collection and usage is crucial, and users must be informed about the potential for AI to generate biased or misleading information (Silva et al., 2023). Ongoing research and development are necessary to mitigate these risks and ensure the ethical and responsible use of AI in applied linguistics.

This strategy emphasizes a balanced approach to AI integration in applied linguistics. By prioritizing ethical considerations, leveraging AI’s strengths for educational and research advancements, and proactively addressing its weaknesses and threats, we can harness the transformative potential of AI while mitigating its risks. This strategy calls for collaboration among researchers, educators, policymakers, and technology developers to ensure the responsible and beneficial integration of AI in the field of applied linguistics.

Limitations

While this study provides a comprehensive SWOT analysis of the integration of generative AI in applied linguistics, several limitations should be acknowledged. The sample of studies was limited to those indexed in Scopus and Web of Science, potentially missing relevant literature from other sources. The qualitative nature of the analysis introduces the potential for subjective interpretation and bias, despite efforts to cross-verify findings. The rapidly evolving nature of AI technology also poses challenges for the timeliness and applicability of the findings, necessitating continuous updates. Additionally, the SWOT framework does not quantify the relative importance of the identified factors, suggesting a need for future research using quantitative methods. Lastly, the study did not delve into specific subfields within applied linguistics, indicating a direction for future research to explore unique implications in areas like language teaching, acquisition, and disorders. Despite these limitations, the study lays a robust foundation for understanding the impacts of generative AI in applied linguistics and offers valuable insights for stakeholders.

Theoretical implications

The integration of AI, particularly through LLMs like ChatGPT and Gemini, into the field of applied linguistics presents significant theoretical implications. The SWOT analysis conducted in this study reveals how AI can transform linguistic theories and frameworks by providing new tools and methodologies for analysing language use and development. For instance, AI’s ability to measure semantic similarity and linguistic patterns aligns closely with cognitive and sociolinguistic theories, offering deeper insights into language processing and social interactions. Additionally, the use of AI in language education supports constructivist and communicative language teaching theories by enabling more personalized and interactive learning experiences. These theoretical advancements can lead to the refinement of existing models and the development of new theories that better account for the complexities of language learning and use in the digital age.

Practical implications

From a practical perspective, the findings of this SWOT analysis suggest several actionable strategies for leveraging AI in applied linguistics to enhance educational and research practices. Educational institutions can utilize AI tools to create more efficient and accessible language learning resources, thereby reducing the workload on educators, and providing students with personalized learning experiences. The development of AI-driven platforms for automatic generation of educational materials, such as vocabulary lists and writing prompts, can streamline classroom instruction and support differentiated teaching approaches. Moreover, the integration of AI in research methodologies can enhance data collection and analysis processes, enabling more robust and scalable linguistic studies. However, it is crucial to address the identified weaknesses and threats, such as ensuring the accuracy of AI-generated content and establishing ethical guidelines to prevent misuse. By implementing these practical strategies, stakeholders in the field can maximize the benefits of AI while mitigating potential risks.