Artificial Intelligence-Based ESP (English For Specific Purposes) Tutor Applications For Science And Engineering Undergraduate Students In Saudi Arabia

While the emergence of artificial intelligence brings changes to the domain of academic assistance, the presence of academic communication continuesin the contexts of foreign language learning, information studies, and human computer interaction. This work considers the usage of AI-driven computing applications used by university students for their assignments related to English for Specific Purposes in science and engineering courses at Saudi higher education establishments. Though the emphasis is made on scientific studies focused on advances in terms of better searching algorithms and language models, no attention is given to chatbot systems that provide students with generic tips on how to write essays. The purpose of this study is to discuss important features involved in designing, teaching, evaluating, and the contextual application of AI tutoring tools. Academic journals and papers written within the scope of interest from 2020 until 2025 were obtained from scholarly indexes. Next, the sources were subject to thematic narrative analysis. Thus, even though the capability of open-ended text production seems promising, computing designs based on fine-tuning language models, embeddings, vector searches, segmenting/indexing approaches, hybrid searches, re-ranking, and prompt handling appear the most reliable way to build an AI tutor in terms of contextualized comments on grammar, coherence, argumentative writing, citation, and writing style. Moreover, it can be stated that hallucination prevention, appropriate source referencing, fairness, security, and Arabic-English dual language tutoring are key principles for building AI tutors. However, as observed from the literature review, there are still inconsistencies concerning criteria and measuring the assessment of writing skills, autonomy, and management.