Tato práce se zabývá možnostmi použití ChatGPT při tvorbě textů určených pro výuku čtení s porozuměním v angličtině jako cizím jazyce. Cílem výzkumu je posoudit účinnost ChatGPT při generování výukových materiálů odpovídajících vzdělávacím cílům pro danou úroveň jazykové znalosti. Analyzujeme lingvistickou strukturu textů generovaných ChatGPT v reakci prompty, které odpovídají deskriptorům CEFR. Tato práce srovnává texty vytvořené pomocí ChatGPT s texty získanými z webu British Council. Metodou je detailní analýza těla textů s použitím volně dostupných lingvistických nástrojů VocabProfiler, TAASSC a T.E.R.A. Tento přístup nám umožňuje zhodnotit rozmanitost slovní zásoby, syntaktickou složitost a čitelnost textů generovaných ChatGPT. Vzhledem k možnostem umělé inteligence v oblasti personalizace výuky jazyka přináší naše studie důležité poznatky pro další výzkum v této oblasti jazykového vzdělávání.
Anotace v angličtině
This thesis explores the utilization of ChatGPT in creating texts tailored for teaching reading comprehension in English as a foreign language. By examining the linguistic makeup of each text generated by ChatGPT in response to prompts at different proficiency levels, the research aims to evaluate the model's efficacy in aiding educational objectives for a varied student population. The thesis compares texts produced by ChatGPT to texts adapted from the British Council. The study adopts a corpus analysis approach, analyzing the texts generated by ChatGPT using linguistic tools VocabProfiler, TAASSC, and T.E.R.A. yielded insights into vocabulary diversity, syntactic complexity, and text readability. Given the potential of AI to enable personalized learning experiences, understanding ChatGPT's efficacy in this context holds significant implications for language education.
Klíčová slova
Generativní umělá inteligence, Výuka anglického jazyka, Nástroje pro výuku jazyků s využitím umělé inteligence, design promptů, Čtení s porozuměním, AI ve školství, pedagogický výzkum, Směrnice pro AI ve vzdělávání
Klíčová slova v angličtině
Generative Artificial Intelligence, English Language Teaching, AI-Powered Language Teaching Tools, Prompt Engineering, Reading Comprehension, Education with AI, Pedagogical Research, AI in Education Policy
Rozsah průvodní práce
128
Jazyk
AN
Anotace
Tato práce se zabývá možnostmi použití ChatGPT při tvorbě textů určených pro výuku čtení s porozuměním v angličtině jako cizím jazyce. Cílem výzkumu je posoudit účinnost ChatGPT při generování výukových materiálů odpovídajících vzdělávacím cílům pro danou úroveň jazykové znalosti. Analyzujeme lingvistickou strukturu textů generovaných ChatGPT v reakci prompty, které odpovídají deskriptorům CEFR. Tato práce srovnává texty vytvořené pomocí ChatGPT s texty získanými z webu British Council. Metodou je detailní analýza těla textů s použitím volně dostupných lingvistických nástrojů VocabProfiler, TAASSC a T.E.R.A. Tento přístup nám umožňuje zhodnotit rozmanitost slovní zásoby, syntaktickou složitost a čitelnost textů generovaných ChatGPT. Vzhledem k možnostem umělé inteligence v oblasti personalizace výuky jazyka přináší naše studie důležité poznatky pro další výzkum v této oblasti jazykového vzdělávání.
Anotace v angličtině
This thesis explores the utilization of ChatGPT in creating texts tailored for teaching reading comprehension in English as a foreign language. By examining the linguistic makeup of each text generated by ChatGPT in response to prompts at different proficiency levels, the research aims to evaluate the model's efficacy in aiding educational objectives for a varied student population. The thesis compares texts produced by ChatGPT to texts adapted from the British Council. The study adopts a corpus analysis approach, analyzing the texts generated by ChatGPT using linguistic tools VocabProfiler, TAASSC, and T.E.R.A. yielded insights into vocabulary diversity, syntactic complexity, and text readability. Given the potential of AI to enable personalized learning experiences, understanding ChatGPT's efficacy in this context holds significant implications for language education.
Klíčová slova
Generativní umělá inteligence, Výuka anglického jazyka, Nástroje pro výuku jazyků s využitím umělé inteligence, design promptů, Čtení s porozuměním, AI ve školství, pedagogický výzkum, Směrnice pro AI ve vzdělávání
Klíčová slova v angličtině
Generative Artificial Intelligence, English Language Teaching, AI-Powered Language Teaching Tools, Prompt Engineering, Reading Comprehension, Education with AI, Pedagogical Research, AI in Education Policy
Zásady pro vypracování
Despite AI's development potential easing the burden on language teachers, analysis of ChatGPT outputs and their alignment to Common European Framework for Reference in teaching Languages is necessary to assess its potential. This thesis systematically examines ChatGPT products highlighting the role of prompt engineering in its text generation capability and offers a comprehensive analysis of ChatGPT outputs for teaching reading comprehension and their efficacy assessment, aiming to provide concise insights and guidelines for effective AI integration in language instruction.
Our goal is to explore whether and if so, how specifically can generative AI be used for production of texts and other teaching implications in teaching reading comprehension. The thesis will assess the suitability of using AI-generated teaching aids and analyze texts produced by ChatGPT on levels A1-B2 according to Common European Framework of Reference for teaching languages. The goal of this thesis is to suggest and discuss potential use of generative AI in English Language Teaching.
The hypothesis is: Under the condition of receiving carefully formulated prompts and instructions, ChatGPT can serve as a valuable resource for English educators. It is anticipated to exhibit effectiveness in the creation of instructional materials specifically adapted to discrete levels of English reading proficiency, ranging from A1 to B2. The verification of this proposition will be undertaken through the construction of a corpus comprised of texts generated by ChatGPT. This compilation will undergo evaluations encompassing various means of measuring lexical and grammatical diversity to authenticate its practical applicability within educational discourse. As a reference corpus, texts adopted from the website of the British Council reading comprehension will be used. The reference corpus will also be divided into 4 parts each corresponding to the language proficiency level.
The thesis attempts to answer the following research questions.
How and in what ways does the linguistic composition of each corpus, generated by ChatGPT in response to specific proficiency-level prompts, reflect the effectiveness of the model in facilitating educational outcomes for diverse learners?
To what extent does ChatGPT demonstrate the ability to generate linguistically suitable texts and adhere to the linguistically suitable patterns for learners across various levels of English proficiency when provided with specific crafted instructional prompts?
What insights can educators derive from the use of ChatGPT in this thesis, and how can these findings aid future possible applications of artificial intelligence to improve reading comprehension instruction in teaching discourse?
Methodology involves extracting reading comprehension descriptors from the Common European Framework for Reference. Subsequently, I formulate a comprehensive instructional prompt tailored for ChatGPT in order to customize its responses. Following this, I delineate prompts specific to each targeted material, ensuring alignment with the anticipated proficiency levels students should attain. Employing these prompts, I instruct ChatGPT to generate sample texts, thereby establishing distinct corpora for each proficiency level. These corpora are subsequently subjected to thorough corpus analysis, the generated corpora are scrutinized to identify patterns and nuances in linguistic output across diverse proficiency levels. This analysis encompasses n-gram analysis, measurement of lexical diversity and density, examination of complexity in discourse markers, and the application of syntactic complexity measures.
Zásady pro vypracování
Despite AI's development potential easing the burden on language teachers, analysis of ChatGPT outputs and their alignment to Common European Framework for Reference in teaching Languages is necessary to assess its potential. This thesis systematically examines ChatGPT products highlighting the role of prompt engineering in its text generation capability and offers a comprehensive analysis of ChatGPT outputs for teaching reading comprehension and their efficacy assessment, aiming to provide concise insights and guidelines for effective AI integration in language instruction.
Our goal is to explore whether and if so, how specifically can generative AI be used for production of texts and other teaching implications in teaching reading comprehension. The thesis will assess the suitability of using AI-generated teaching aids and analyze texts produced by ChatGPT on levels A1-B2 according to Common European Framework of Reference for teaching languages. The goal of this thesis is to suggest and discuss potential use of generative AI in English Language Teaching.
The hypothesis is: Under the condition of receiving carefully formulated prompts and instructions, ChatGPT can serve as a valuable resource for English educators. It is anticipated to exhibit effectiveness in the creation of instructional materials specifically adapted to discrete levels of English reading proficiency, ranging from A1 to B2. The verification of this proposition will be undertaken through the construction of a corpus comprised of texts generated by ChatGPT. This compilation will undergo evaluations encompassing various means of measuring lexical and grammatical diversity to authenticate its practical applicability within educational discourse. As a reference corpus, texts adopted from the website of the British Council reading comprehension will be used. The reference corpus will also be divided into 4 parts each corresponding to the language proficiency level.
The thesis attempts to answer the following research questions.
How and in what ways does the linguistic composition of each corpus, generated by ChatGPT in response to specific proficiency-level prompts, reflect the effectiveness of the model in facilitating educational outcomes for diverse learners?
To what extent does ChatGPT demonstrate the ability to generate linguistically suitable texts and adhere to the linguistically suitable patterns for learners across various levels of English proficiency when provided with specific crafted instructional prompts?
What insights can educators derive from the use of ChatGPT in this thesis, and how can these findings aid future possible applications of artificial intelligence to improve reading comprehension instruction in teaching discourse?
Methodology involves extracting reading comprehension descriptors from the Common European Framework for Reference. Subsequently, I formulate a comprehensive instructional prompt tailored for ChatGPT in order to customize its responses. Following this, I delineate prompts specific to each targeted material, ensuring alignment with the anticipated proficiency levels students should attain. Employing these prompts, I instruct ChatGPT to generate sample texts, thereby establishing distinct corpora for each proficiency level. These corpora are subsequently subjected to thorough corpus analysis, the generated corpora are scrutinized to identify patterns and nuances in linguistic output across diverse proficiency levels. This analysis encompasses n-gram analysis, measurement of lexical diversity and density, examination of complexity in discourse markers, and the application of syntactic complexity measures.
Seznam doporučené literatury
Ali, Z. (2020, February). Artificial intelligence (AI): A review of its uses in language teaching and learning. In IOP Conference Series: Materials Science and Engineering (Vol. 769, No. 1, p. 012043). IOP Publishing.
Doulík, P. (2016). Vybrané základy metodologie pedagogického výzkumu: (se cvičeními). Univerzita J.E. Purkyně v Ústí nad Labem.
Harmer, J. (2007). The practice of English language teaching (4th edition.). Longman.
Homels, W., Bialik, M., Fadel, Ch. (2019). Artificial Intelligence in Education. Promise and Implications for Teaching and Learning. Pearson
Gurung, R. A., & Schwartz, B. M. (2011). Optimizing teaching and learning: Practicing pedagogical research. John Wiley & Sons.
Newby, D. (2007). European Portfolio for Student Teachers of Languages: A reflection tool for language teacher education. Council of Europe.
Seznam doporučené literatury
Ali, Z. (2020, February). Artificial intelligence (AI): A review of its uses in language teaching and learning. In IOP Conference Series: Materials Science and Engineering (Vol. 769, No. 1, p. 012043). IOP Publishing.
Doulík, P. (2016). Vybrané základy metodologie pedagogického výzkumu: (se cvičeními). Univerzita J.E. Purkyně v Ústí nad Labem.
Harmer, J. (2007). The practice of English language teaching (4th edition.). Longman.
Homels, W., Bialik, M., Fadel, Ch. (2019). Artificial Intelligence in Education. Promise and Implications for Teaching and Learning. Pearson
Gurung, R. A., & Schwartz, B. M. (2011). Optimizing teaching and learning: Practicing pedagogical research. John Wiley & Sons.
Newby, D. (2007). European Portfolio for Student Teachers of Languages: A reflection tool for language teacher education. Council of Europe.