Assessing the Accuracy of ChatGPT in Translating English Journalistic Texts into Arabic

تقييم دقة ChatGPT في ترجمة النصوص الصحفية من الإنجليزية الى العربية

محمد عسكر مصري علي¹     عبد الرحمن احمد  القزاز²

¹قسم الترجمة، كلية الآداب، جامعة الموصل، الموص، العراق. ² قسم الترجمة، كلية الآداب، جامعة الموصل، العراق

المستخلص

يسعى هذا البحث الى تقييم دقة نوعية ترجمة خدمة ChatGPT لمعرفة مدى تطبيقها العملي وكفاءتها ودلالة عملها. حيث أن أنموذج لارسون (1984) المتبنى يقوم بتقييم أخطاء دقة الترجمة وفقاً لـ: الدقة والوضوح والطبيعية. فمن خلال معيار الدقة (كمعيار فرعي (لتحليل الأخطاء، تهدف هذه الدراسة الكشف عن العوامل المؤثرة في خدمة ترجمة ChatGPT وتغطيتها عن طريق تحديد الأنواع المتكررة للإخطاء في مخرجات الترجمة الالية من خلال استخدام أنموذج لارسون (معيار الدقة). وتفترض هذه الدراسة أن تزويد المعلومات الخاطئة يعتبر من أكثر الأخطاء شيوعاً طبقاً لمعيار الدقة في ترجمة النصوص الصحفية من الإنكليزية الى العربية. أضافةً الى ذلك، أن ترجمة وتقييم النصوص الصحفية تعد أقل تحدياً فيما لو كانت هنالك معرفة لغوية وغير لغوية مناسبة لمبادئ أسلوب الكتابة الصحفية. حيث أن النتائج تظهر بان مخرجات ترجمة  ChatGPTنوعاً ما لا تزال تحتاج الى تدخل أنساني لتعديل وترتيب الأخطاء على الرغم من كون الترجمة الالية قد قامت بتحسين من كفاءة ترجمتها. وبناءً على تحليل أخطاء دقة الترجمة، فإن نقل معلومة خاطئة أو الحذف والإضافة الغير مبررين من الممكن ان يخلق مشكلة في ترجمة ChatGPT للنصوص الصحفية من اللغة الإنكليزية الى اللغة العربية مما يعني أن تطبيق الارشادات الصحيحة الأساسية مهمة في تنقيح الترجمة ولذلك يتوجب على المترجمين المنقحين أن يكونوا على دراية كافية بالأخطاء المتكررة.

الكلمات المفتاحية: تشات جي بي تي؛ الترجمة الالية؛ الدقة؛ تقييم نوعية الترجمة؛ الذكاء الاصطناعي

1. Introduction

     Since ChatGPT has become a vital source of knowledge across distinct domains, its significance continues to grow. Additionally, numerous studies have shown that at the core of translation quality assessment lies the prevailing questions of reliability, practicality, and validity, three central concepts in the sub-branch of translation studies. Therefore, many scholars have sought to quantify the time and effort consumed during the translation process. The current paper sheds light on the assessment of ChatGPT's translation production by applying Larson’s Translation Quality Assessment model to three journalistic texts.

     In brief, this study considers Larson’s (1984, p. 53) (1) criteria (Accuracy as a significant standard factor). Larson's model is applied to evaluate the translation quality of a diverse range of pragmatic texts, including journalistic texts, a genre the study closely investigates. While taking into account the linguistic, social, and cultural contexts in which it occurs within an interdisciplinary structure of text analysis that reckons with linguistic and extra-linguistic factors. Larson’s model also examines news translation along three distinct yet essential axes: the unjustified omission, the unjustified addition, and the provision of incorrect information. The translation goes beyond the linguistic limit of lexical recapturing to become a dynamic, situated act of communication.

2. Literature Review

     Over the last few decades, with the emergence of digital computers, (Koehn, 2010) (2) states that machines have taken on many human tasks, including those once believed to require thinking and intelligence. Translating between languages is one of such tasks. In recent years, there has been a surge of interest in post-translation editing and in free, high-quality software, driven by improvements in the quality of Machine Translation (MT) output. Post-Editing (PE) is the oldest form of human-machine cooperation in translation, and it remains a widespread practice for operational MT systems. The usefulness of MT in the PE scene (Koby, 2001, p. 51) may be due to its high reliability. Hence, PE may end up taking even more time and effort than discarding the MT and translating from the very beginning, unless the raw MT is of poor quality.

2.1 Beneficial results of working with MT:

     Harold Somers (2003, p. 58) (4) postulates that the purpose of sustaining a glossary of any type is to stimulate terminological regularity and to hinder translators from having to duplicate their research each time they start a new translation project. It is unnecessary to use specialized software to maintain a glossary — several translators have long used card indexes or word processors to compile terminology records. These results can be seen in the quality, speed, flexibility, and shareability of information in the translation process and final output.  

2.2 Failure of Machine Translation:

     Machine Translation (MT)  Mahmoud (2023, pp. 160-166) (5) has proved that many challenges and limitations hinder the smooth flow of the final translation process; however, its variety of benefits and constant advancement at the expense of human traditional translation. From the perspective of translation knowledge, the following are some common limitations of access-free internet-based translation services: literal translation, grammatical mistakes, conversational errors, spelling errors, social references, and irregular writing.

2.3 Machine Translation Systems:

     Almahasees (2022:4-2) (6) signifies that the development of MT can be traced back to the advancement of NLP research and implementations, AI development, the media revolution, and the requirement for ensuring international communication among people speaking various languages. Consequently, considerable MT systems have emerged and developed rapidly. Many systems are widely used, including Google, Microsoft Translator, Sakhr, and ChatGPT. Each of these systems supports Fully Automatic Translation (FAT) for English and Arabic across platforms: mobile devices, online, desktop, speech translation, offline, and voice.

2.4 ChatGPT Translation Tool

     In his book, Sarrion (2023) personifies ChatGPT, a large-scale natural language processing (NLP) approach developed by OpenAI. NLP attempts to enable computers to interact and understand human language. Since ChatGPT uses machine learning to generate text based on the user's input, this can be important for human communication and interaction. Therefore, ChatGPT can be used to spark conversations with humans. ChatGPT, in this domain, is a pre-trained NLP model that has become widely used recently. The GPT (abbreviation of “Generative Pre-trained Transformer" utilizes complicated artificial neural networks.

2.5 Larson`s Model of Translation Quality Assessment:

     The evaluation step (1984:529), which follows the initial draft, includes threefold matters: (1) accuracy, (2) clearness, and (3) naturalness. These threefold are to evaluate the translation output that Larson evokes four questions to be answered on the various aspects of evaluation: (a) Does the translation transfer the same meaning as the SL? (b) Does the TL audience understand it clearly? (c) Is the form of the translation easily read and natural in the recipient's language grammar and style?

2.6 Accuracy Criteria

     Larson (1984:529) (1) postulates accuracy (or as she called a comparison test) as the procedure of checking out the translation product at several points by a careful comparison of the TT with the ST to avoid unjustified omission of some pieces of information. In addition, some mistakes in the analysis of the ST or the transfer process led to different meaning results. Since it is easy to make mistakes at any stage of the translation process, a precise comparison with the ST will be required several times. Therefore, the essential purpose of the accuracy (comparison) test is to check for equivalence of information content and to be sure that all information is involved, nothing omitted, nothing added, and nothing different within the TT.

3. Research Methodology

Research Design: The present paper adopts a qualitative and quantitative methodology by analyzing written English journalistic texts translated into Arabic by ChatGPT through Larson's "Translation Quality Assessment (TQA) Model." Three English news texts (articles) are carefully selected for a comprehensive analysis, discussion, and application of the model to verify the hypotheses with findings and conclusions.

3.1 Text (A)

Data Analysis: Text (A) is taken from (The Guardian Weekly) published on (Mon 3 Feb 2025 15.48 GMT). This edition features a weekly interview conducted by Simon Bland, the editor of The Guardian Weekly, with Jim Carrey, a prominent American-Canadian actor, and Jeff Daniels, an American actor. The discussion highlights their performances in the awkward scenes in the “Dumb and Dumber” movie in 1994. The interview offers insights into their collaborative creative process and the significant role played by Daniels in this influential cinematic work.

Data Analysis Procedures:

Para. (1), Sent. (1): ‘Jim Carrey told me to “go all the way” with the laxative toilet scene’: Jeff Daniels on Dumb and Dumber.

قال لي جيم كاري "أنطلق إلى الاخر" في مشهد المرحاض المليء بالملين": جيف دانييلز عن فيلم الغبي والأغبى.

Para. (2), Sent. (6): I remember thinking: “It’s either going to be such a bad career move I’ll never work again, or it just might be funny.”

                 أتذكر أنني فكرت: "إما أن تكون هذه خطوة سيئة في مسيرتي تجعلني لا أعمل مرة أخرى، أو قد تكون مضحكة.

Para. (3), Sent. (6): Then I’d towel my hair dry and not comb it. And that was it.

    ثم قمت بتجفيف شعري بالمنشفة دون تمشيطه.

Para. (4), Sent. (1): The first morning of shooting had me and Jim up against each other on a little scooter going down mountains.

                في صباح اليوم الأول من التصوير، كنا أنا وجيم على دراجة صغيرة ننزل من الجبال.

Para. (5), Sent. (2): It’s one of the few times where you can see me break into laughter

                إنها واحدة من المرات القليلة التي يمكنك فيها رؤيتي أكسر إلى الضحك

Para. (6), Sent. (5): I said: “Just cold feet.

  قلت: "فقط أشعر بالبرودة.

Para. (7), Sent. (5): I still have a scrapbook of 200 newspapers panning the movie and wishing it never existed.

                لديّ كتيب من 200 جريدة تنتقد الفيلم وتتمنى أنه لم يكن موجودًا.

Analytical Framework

Discussion: The (Para.1, Sent.1) translation, in terms of accuracy criterion, represents unjustified addition and wrong information, although it implies no omission. The translation of (‘Jim Carrey told me to “go all the way” with the laxative toilet scene’: Jeff Daniels on Dumb and Dumber) into (قال لي جيم كاري "أنطلق إلى الاخر" في مشهد المرحاض المليء بالملين": جيف دانييلز عن فيلم الغبي والأغبى) reveals inaccurate translation. Firstly, (المليء) is an unjustified addition because it distorts the meaning. Secondly, the translation of (the laxative toilet scene) into (مشهد المرحاض المليء بالملين) does not fully convey the meaning, and it indicates a distorted meaning because the scene uses a laxative drug by the actor, and the toilet itself is not a (laxative), which is logically inaccurate. A more idiomatic and precise translation could reflect the logical reason between the laxative and the toilet. A simple and accurate translation would be (في مشهد المرحاض بعد تناول دواء الملين), or (مشهد المرحاض بعد تناول الدواء المسهل). On the contrary, in the translation (Jeff Daniels on Dumb and Dumber) into (جيف دانييلز يتحدث عن فيلم الغبي و الاغبى)

     ChatGPT translation in (Para. 2, Sent. 6) shows an essential sign of a lack of accuracy. The TT provides a distorted meaning because (مضحكة) has polysemous references and may refer to something as ridiculous or interesting rather than funny. Furthermore, ChatGPT translation omits the lexical items (تجربة) or (خطوة) to emphasize TT accuracy.

     Once again, ChatGPT seems to focus less on the accuracy criterion in the translation of (Para.3, Sent.6). Grammatically, it loses the habitual nuances because of the choice of the verb where the ST (I’d towel my hair dry and not comb it) implies a habitual action in the past, while the TT (قمت بتجفيف شعري بالمنشفة دون تمشيطه) indicates a one-time action. The translation should reflect a habitual aspect of the TT to convey the intended meaning accurately. A more precise equivalent that maintains the intended meaning would be if we replace (قمت بتجفيف) with (كنت أجفف).

     The translation of (Para. 4, Sent. 1) lacks the accuracy criterion mark, since translating (had Jim and me up against each other) as (كنا أنا وجيم) is inaccurate because it implies an external force and places them in a specific situation or position. Semantically, the translation of (we were) into (كنا) does not fully capture the intended nuance. A better translation would be (وجدت نفسي أنا وجيم) because it powerfully conveys the meaning of context.

     The translation in (Para. 5, Sent. 2) also exhibits a missed mark of accuracy because the verbal sentence (أكسر الى الضحك) is inaccurate and forces the reader to stop reading to think about the context since the TT does not fully convey the sense of suddenness or spontaneity in the way the ST does. A more vivid and accurate translation would be (أنفجر ضاحكاً) or (أضحك فجأةً).

     (Para.6, Sent.5) Translation also lacks an accuracy criterion because the translation of (I said: Just cold feet) into (قلت: فقط أشعر بالبرودة) fails to capture the intended meaning of ST. ChatGPT produces wrong information due to incorrect lexical choices. A more accurate and precise translation would be (قلت: إنه مجرد توتر), which conveys mild nervousness.

     The translation of (Para. 7, Sent. 5) points out signs of a less accurate translation characterized by lexical and contextual inaccuracies that distort the essence of the ST. Lexically, the choices of (كتيب), (تنتقد), and (تتمنى أنه لم يكن موجوداً) are inaccurate equivalents for (scrapbook), (panning), and (wishing it never existed) since they provide wrong information and misinterpret the meaning. A more accurate translation would be (لا زلت أحتفظ بألبوم يحتوي على 200 مقالة من مختلف الصحف تهاجم الفيلم وتصفه بأنه لا يستحق أن يكون موجودًا). This translation would not only convey the core message but also preserve the source sensation with naturalness.

3.2 Text (B)

Data Analysis: Text (B) is taken from the (CNN) published on (3:48 AM EST, Fri March 7, 2025). This edition features a compelling military and political analysis conducted by Brad Lendon, the Senior News Desk Reporter, and Global Military Affairs for CNN International. The text delves into the potential effects of the US's reducing contribution, especially during the presidency of Donald Trump. The text also highlights that if European countries unite and invest in the right military equipment, Europe can still be a significant deterrent to Russia. The text points out the historical role of the US in NATO, particularly during the Cold War and the Balkans conflicts, and points to recent developments that may signal a change in intercontinental solidarity.

Data Analysis Procedures

Headline: Can NATO survive without the United States?

هل يمكن لحلف الناتو البقاء دون الولايات المتحدة؟

Para. (2), Sent. (2): and recent comments casting doubt over whether he would defend NATO allies “if they don’t pay” have all forced European leaders to start thinking the previously unthinkable

إلى جانب تصريحاته الأخيرة التي شكك فيها فيما إذا كان سيدافع عن حلفاء الناتو "إذا لم يدفعوا"، إلى إجبار القادة الأوروبيين على التفكير في أمر كان يعتبر سابقًا غير وارد

Para. (3), Sent. (2): It also has the wealth and technological know-how to defend itself without the US, analysts say.

كما أن لديه الثروة والمعرفة التكنولوجية اللازمة للدفاع عن نفسه دون الحاجة إلى الولايات المتحدة، وفقًا لما يقوله المحللون.

Para. (4), Sent. (1): The US and Germany are the biggest contributors to NATO’s military budget, civil budget and security investment program, at almost 16% each, followed by the UK at 11% and France at 10%, a NATO fact sheet says.

                تُعد الولايات المتحدة وألمانيا أكبر المساهمين في ميزانية الناتو العسكرية، والميزانية المدنية، وبرنامج الاستثمار الأمني، بنسبة تقارب 16% لكل منهما، تليهما المملكة المتحدة بنسبة 11%، وفرنسا بنسبة 10%، وفقًا لورقة معلومات صادرة عن الناتو.

Analytical Framework

Discussion: The (Headline) translation fails to meet Larson`s accuracy criteria. The translation of (Can NATO survive without the United States? ") into (هل يمكن لحلف الناتو البقاء دون الولايات المتحدة؟) reveals inadequate lexical choices that lessen the essence of the ST and misinterpret the intended meaning. For instance, (Can) in the ST refers to NATO`s capability rather than the possibility of survival. Furthermore, (survive), as the contextual usage, indicates maintaining strength despite challenges, whereas (البقاء) means (remaining or not disappearing) since it conveys a passive state of existence. A more accurate and precise translation would be (هل يستطيع حلف الناتو الصمود دون الولايات المتحدة؟), which captures the essence of resilience that the original headline represents.

     The translation in (Para.2, Sent.2) fails to meet the accuracy criterion. Since the translation of (and recent comments casting doubt over) into (إلى جانب تصريحاته الأخيرة التي شكك فيها) introduces additional information like (الى جانب) due to the incorrect lexical choice and (شكك فيها) also provides wrong information resulting in the misinterpretation of the intended meaning of the ST.  Notably, it still demonstrates NO omission. A more accurate and nuanced translation would be (وتصريحاته الاخيرة أثارت التساؤلات حول), which better conveys the intended meaning. Additionally, the structural choices made by ChatGPT lead to inaccuracies that misinterpret the ST and distort its meaning. For instance, (فيما إذا كان) and (إلى إجبار) exemplify an unnecessarily complicated structure and a lack of a direct causal link.

     The translation in (Para.3 Sent.2) fails to meet the accuracy criterion because it does not fully capture the intended meaning of the ST. Since the ChatGPT translation of (It also has the wealth and technological knowhow to defend itself without the US) into (كما أن لديه الثروة والمعرفة التكنولوجية اللازمة للدفاع عن نفسه دون الحاجة إلى الولايات المتحدة) slightly implies a distorted meaning due to the inaccurate lexical choice. The phrase (المعرفة التكنلوجية) refers to the theoretical understanding rather than practical expertise. A more delicate and accurate translation would be (المهارة التكنلوجية) or (الخبرة التقنية). Additionally, ChatGPT falls short due to the unnecessary and unjustified addition of (technological know-how to defend itself) into (التكنولوجية اللازمة للدفاع عن نفسه). This extra lexeme (اللازمة) is not present in the ST, and it violates the intended meaning. To avoid the accuracy error; it would be more accurate to omit the lexeme.

     Once again, the accuracy criterion misses the mark in the translation of (Para.4, Sent.1) due to the inappropriate lexical choices. The expressions (الميزانية المدنية/ Civil budget) and (ورقة معلومات/ Fact sheet) reflect an inaccurate translation since the ST addresses the military issues rather than humanitarian ones. Improved equivalents as (الميزانية الادارية) instead of (الميزانية المدنية) and (وثيقة حقائق) instead of (ورقة معلومات) might result in more accurate translation.

3.2 Text (B)

Data Analysis

      Text (C) is taken from (Michigan Daily), published on (October 27, 2024). This edition features a daily contribution by Pauline Kim, a skilled contributor to The Michigan Daily, the independent student newspaper of the University of Michigan. Kim concentrates on literature and personal essays that explore the intersection of reading and self-reflection.

Data Analysis Procedures

Para. (1), Sent. (1):            When I was in 7th grade, I lied and said I liked “Pride and Prejudice.”

عندما كنت في الصف السابع، كذبت وقلت إنني أحب "فخر واحتقار".

Para. (1), Sent. (4): For a long time afterward, I kept a mental rant stored in my back pocket about how classics are boring, stuffy and way too long.

                لفترة طويلة بعد ذلك، احتفظت بصراع ذهني مخزن في جيبي حول كيف أن الكتب الكلاسيكية مملة، ومتصلبة وطويلة جدًا.

Para. (3), Sent. (2): They change the way you look at life.

إنها تغير الطريقة التي تنظر بها إلى الحياة.

Para. (3), Sent. (3): In a world of muddled browns and tactile reds, where we judge a person by their political leanings, there’s always someone dying somewhere in the world and everyone is anxious to not offend anyone else — a headache by any standard — classics are like a puff of light blue smoke.

في عالم مليء بالبني المشوش والأحمر الملموس، حيث نحكم على شخص ما بناءً على توجهاته السياسية، وهناك دائمًا شخص يموت في مكان ما في العالم والجميع قلق من أن يسيء إلى الآخرين — إنها صداع بمعايير أي شخص — الكلاسيكيات هي مثل نفخة من الدخان الأزرق الفاتح.

Para. (4), Sent. (2): It’s a pure, unedited glimpse into an older generation, people who were fundamentally just like us, albeit in a completely different context.

                إنها لمحة نقية وغير معدلة عن جيل قديم، أشخاص كانوا في الأساس مثلنا، رغم أنهم كانوا في سياق مختلف تمامًا

Para. (4), Sent. (4): We so easily accept that so much of human belief and behavior is a result of the environment we were born into, so it’s not that hard to imagine that some future generation will stroll down the block and condemn everything we ever stood for as backward, ridiculous and possibly even pure evil.

نحن نقبل بسهولة أن الكثير من المعتقدات والسلوكيات البشرية هي نتيجة للبيئة التي وُلدنا فيها، لذا ليس من الصعب أن نتخيل أن جيلًا مستقبليًا سوف يمر في الشارع ويدين كل ما وقفنا من أجله باعتباره متخلفًا، سخيفًا وربما حتى شريرًا تمامًا.

Para. (5), Sent. (1):            For all our generational and cultural differences, certain values and struggles still ring true.

                على الرغم من اختلافاتنا الجيلية والثقافية، فإن بعض القيم والنضالات لا تزال تتردد صداها.

Para. (5), Sent. (5): When Oscar Wilde writes, “Women defend themselves by attacking, just as they attack by sudden and strange surrenders.”

عندما يكتب "أوسكار وايلد"، "الدفاع عن النفس عند النساء يكون عن طريق الهجوم، تمامًا كما يهاجمن عبر استسلامات مفاجئة وغريبة."

Para. (6), Sent. (1):            What we often dismiss as obsolete is still very much human, and though the societal definitions of “good” and “bad” shift every couple of decades, we have so much to learn from the people who came before us.     

ما نرفضه عادة كقديمة لا يزال بشريًا جدًا، وعلى الرغم من أن التعريفات الاجتماعية لـ "الخير" و"الشر" تتغير كل بضع عقود، فإن لدينا الكثير لنتعلمه من الأشخاص الذين عاشوا قبلنا.

Para. (7), Sent. (1): A funny result of it all is that I feel more connected to humanity as a whole, all 117 billion of us who have ever walked and lived and breathed on this earth.

نتيجة مضحكة لكل ذلك هي أنني أشعر باتصال أكبر بالإنسانية ككل، 117 مليارًا منا الذين ساروا وعاشوا وتنفسوا على هذه الأرض.

Analytical Framework

Discussion: The translation presented in (Para.1, Sent.1) fails to meet the accuracy criterion. Notably, the translation of (I liked Pride and Prejudice) into (إنني أحب فخر واحتقار) exemplifies a misinterpretation meaning due to the inaccurate lexical choices since the (فخر/pride) and (احتقار/ prejudice) do not fully capture the intended meaning of the ST thereby produces incorrect information. A more accurate and delicate translation would be (الكبرياء و التحامل).

     The accuracy criterion loses a mark in (Para.1, Sent.4) due to the incorrect lexical choices. The translation of (For a long time afterward, I kept a mental rant stored in my back pocket about how classics are boring, stuffy, and way too long) into (لفترة طويلة بعد ذلك، احتفظت بصراع ذهني مخزن في جيبي حول كيف أن الكتب الكلاسيكية مملة، ومتصلبة وطويلة جدًا) exemplifies wrong information. ChatGPT choices of (صراع ذهني) and (متصلبة) reveal inaccurate lexical and stylistic selections, resulting in a distorted meaning of the ST. A more precise and delicate translation would be (تذمر ذهني) and (متكلفة).

     The translation in (Para.3, Sent.2) violates the accuracy criterion due to providing wrong information that distorts the intended meaning of the ST. The translation of (They change the way you look at life) into (إنها تغير الطريقة التي تنظر بها إلى الحياة) exhibits a misinterpretation since the pronoun in ST (They) refers to the plural. In contrast, in TT, the pronoun (إنها) refers to the singular, which is inapplicable to the numbers in the ST. To avoid misinterpretation, a better and more delicate translation would be (هي التي تغير نظرتك الى الحياة) or being more specific with the subject (الكتب هي التي تغير نظرتك في الحياة).

     ChatGPT falls into the trap of translating metaphoric expressions, such as in (Para.3, Sent.3), which does not personify the accuracy criterion, stipulating any addition, or wrong information. For instance, the translation of (In a world of muddled browns and tactile reds) into (في عالم مليء بالبني المشوش والأحمر الملموس) reveals a ChatGPT's literal translation since the emotional and symbolic connotation is omitted. Additionally, the translation of (a headache by any standard) into (إنها صداع بمعايير أي شخص) shows an unjustified addition (شخص), which misleads the translation of the intended meaning. Furthermore, the translation of (a puff of light blue smoke) into (نفخة من الدخان الأزرق الفاتح) shows wrong information. The translation of these metaphoric expressions essentially leads to a distorted meaning of the ST. A more metaphoric and accurate rendition would be (في عالمٍ تغشاه الألوان الكئيبة والانفعالات الحادة), (موقف خانق بكل المقاييس), and (كنسمة دخان ازرق باهت) to fully capture the connotation of the ST.

     In contrast, the translation in (Para. 4, Sent. 2) does not meet the accuracy criterion. Since the translation of (It’s a pure, unedited glimpse into an older generation, people who were fundamentally just like us, albeit in a completely different context) into (إنها لمحة نقية وغير معدلة عن جيل قديم، أشخاص كانوا في الأساس مثلنا، رغم أنهم كانوا في سياق مختلف تمامًا) reveals neither any omission nor addition. However, it still provides incorrect information due to inaccurate lexical choices, distorting the ST's meaning. Specifically, the term (unedited) refers to emotional or stylistic connotation, in contrast to (غير معدلة) in TT, which indicates mechanical or technical editing. In the same vein, the phrase (older generation) refers to a relative generational difference rather than (جيل قديم) historically old generation. A more accurate translation would be (أصيل/ genuine), which indicates originality and purity, and (جيل سابق or جيل من الأجيال الماضية), which means a previous generation. (A Dictionary of Modern Written Arabic, p. 20).

     Although the translation in (Para.4, Sent.4) includes neither any omission nor addition. It still shows deviations in lexical choice and idiomatic expressions. The translation of (It’s not that hard to imagine that some future generation will stroll down the block and condemn everything we ever stood for as backward, ridiculous and possibly even pure evil) into (ليس من الصعب أن تتخيل أن جيلاً مستقبلياً سوف يمر في الشارع ويدين كل ما وقفنا من أجله باعتباره متخلفاً, سخيفاً و ربما شريراً تماماً) conveys wrong information due to the imprecise lexical choices, resulting in a distorted meaning of the ST. The phrase (stroll down) implies a reflective, leisurely walk, while (يمر) means merely passing by. A more contextually appropriate translation would be (يتجول في الحي) or (يسير على مهل في الشارع).

     The translation in (Para.5, Sent.1) lacks an accuracy criterion mark due to the inaccurate lexical choice, providing the wrong information due to the literal rendering. The translation of (For all our generational and cultural differences, certain values and struggles still ring true) into (على الرغم من اختلافاتنا الجيلية والثقافية، فإن بعض القيم والنضالات لا تزال تتردد صداها) misrepresents the intended meaning by favoring a literal sound-perception, not the exist in the ST, fails to convey the essence. The idiomatic expression (still rings true) concentrates on the truthfulness or ongoing relevance. While (لا تزال تتردد صداها) focuses on the acoustic echo. A more accurate translation would be (لا تزال حاضرة حتى اليوم) or (لا تزال تحتفظ بصدقها).

     The translation in (Para. 5, Sent. 5) fails to meet the accuracy criterion because of the inaccurate lexical choices, leading to a distorted meaning of the ST. The translation of (Women defend themselves by attacking) into (الدفاع عن النفس عند النساء يكون عن طريق الهجوم) implies wrong information due to the literal translation.

     The lexeme (defend) functions as a lexical verb in ST, referring to psychological and emotional behavior rather than to physical behavior. Besides, the Arabic language typically favors a verbal structure; however, the lexeme (الدفاع) functions as a noun in TT, which is a nominal structure, and it also refers to physical behavior more than the other types. A more accurate and delicate translation would be (تُدافع النساء عن أنفسهن بطريقة الهجوم) or (الهجوم هي احدى طرق الدفاع لدى النساء عن أنفسهن). 

     The translation in (Para.6, Sent.1) violates the accuracy criterion because of the inaccurate lexical choices and structural mismatches, resulting in a distorted meaning. The translation of (What we often dismiss as obsolete) into (ما نرفضه عادة كقديمة) indicates misleading information. Lexically, the lexeme (dismiss) denotes assuming something or someone is unimportant or unworthy of attention. In stark contrast, the lexeme (نرفض) implies throwing away or rejecting the matter as worthless. In addition, the lexeme (obsolete) refers to something outdated or irrelevant. Meanwhile, the lexeme (قديمة) does not necessarily carry the negative or dismissive inference, and it lacks the functional and contextual nuance of (obsolete). On the other hand, the structural mismatching leads to inaccurate translation in that the lexeme (obsolete) refers to an abstract idea or concrete item, which is not clear, while (ما نرفضه كقديمة) denotes a feminine adjective, which does not make any sense in Arabic without a corresponding feminine noun. A more precise translation would be (ما نتجاهله في كثير من الأحيان شيءٌ عفا عليه الزمن).

     Since the translation includes an omission, the accuracy criterion misses the mark of inadequate translation. The omission occurs in the second part of (Para.7, Sent.1), leading to a less accurate output. The translation of (all 117 billion of us who have ever walked) into (117 مليارًا منا الذين ساروا) reveals the omission of the term (ever), which there is no direct equivalent in the TT that captures the same nuance without shifting the structure and potentially providing a misinterpretation. A more precise translation would be (جميع 117 مليار منا الذين ساروا), which more faithfully conveys the intended meaning of the ST.

3.4 Findings

The study has yielded the following findings:

1-     Translation of figurative expressions is insufficient due to its tendency to literal translation or semantic translation rather than communicative one. The total number of translation accuracy errors across all translation texts is (26), out of which text (A) errors make up (7) (28%), text (B) errors make up (6) (24%), and text (C) errors make up (12) (48%). ChatGPT.

2-     The total number of accuracy errors is (26). The highest rate for lack of accuracy in most translations is attribution to providing wrong information, with (19) (73.07%) accuracy errors. Inaccuracy due to omission of information is the second highest rate with (5) (19.23%) of all accuracy errors, and inaccuracy due to addition comes in third with (2) (7.69 %) of all accuracy errors

Table 3: Statistical analysis of accuracy criteria

Table (4) Statistical Text Accuracy Criteria Analysis

4 . Conclusion

1-    ChatGPT Translation of figurative expressions is insufficient due to its tendency to literal translation or semantic translation rather than communicative one. For instance; in text (A), the translation of (‘Jim Carrey told me to “go all the way”) into (قال لي جيم كاري "أنطلق إلى الاخر"), in the text (B), the translation of (Europe is staring down the barrel of a stark new reality) into (تواجه أوروبا واقعًا جديدًا صارخًا), and in the text (C), the translation of (In a world of muddled browns and tactile reds) into (في عالم مليء بالبني المشوش والأحمر الملموس).

2-    Larson’s model substantiates the validity, efficiency, and practicality when applied to assess the Translation Quality Assessment (TQA) of English journalistic texts into Arabic. The application of this model makes the universal criteria of assessment easy, practical and objective; it also sheds the lights on various aspects of texts.

3-    The requirement of human post-editing is necessarily as ChatGPT translation shows repeated issues in lexical choice, idiomaticity and to enhance not only the semantic accuracy but also to ameliorate the naturalness and clarity of the TT according to the TL norms.

5 Recommendations

The study offers a number of recommendations:

1-    For translation pedagogical issues this model can be useful as it offers a clear show case of translation values.

2-    The semantic accuracy assurance confirms that the translated text authentically conveys the ST message, especially idiomatic expressions, metaphors, and cultural nuances by reviewing the TT for any violations if there is any unjustified omission, or addition in contexts. 

Acknowledgement

I wish to articulate my profound and unfeigned gratitude to Alnoor University for bestowing the institutional patronage, intellectual latitude, and material resources indispensable for the successful execution of this scholarly inquiry—generously underwritten. This emblematic support underscores the university’s unwavering commitment to fostering advanced humanistic scholarship.

     My deepest academic reverence is owed to my supervisor, Prof. Dr. Abdulrahman Ahmed Al-Qazzaz, whose sagacious mentorship, incisive intellect, and steadfast encouragement have served as an intellectual lodestar throughout the labyrinthine paths of this research. His discerning feedback and scholarly rigor have profoundly shaped the path and depth of this work.

References

1.Larson M L.  Meaning-Based Translation: A guide to Gross Language Equivalenc (Vol. 2nd Edition). New York. Oxford: University Press of America, New York. Oxford . 1984.

2.Koehn P. Statistical Machine Translation. Cambridge, New York, Melbourne, Madrid, Cape Town, Singapore: Cambridge University Press The Edinburgh Building, Cambridge CB2 8RU, UK. Retrieved from 2010. www.cambridge. org/ 9780521874151

3.Koby H P. Repairing Texts: Empirical Investigations Of Machine Translation Post Editing Processes. The United States of America : The Kent State of University Press, Kent, Ohio, 44242. Retrieved—from. 2001.  http://www. bookmasters. com/ ksu-press.

4.Somers H.  Computers and Translation A translator’s guide. The Netherlands: John Benjamins, Amsterdam·2003.

5.Mahmoud A A. Failure of Machine Translation of English Dysphemistic Expressions into Arabic. Adab Al-Rafidin. 2023;53(95):155-174.

6.Almahasees Z. Analyzing English-Arabic Machine Translation: Google Translate, Microsoft Translator and Sakhr. New York: Routledge, 2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN. 2022.

7. Sarrion E. Exploring the Power of ChatGPT: Applications, Techniques, and Implications. (2023). Paris, France: . Retrieved from https://doi.org /10.1007/978-1-4842-9529-8

APPENDECIES

Text (A)

Dumb and Dumber

‘Jim Carrey told me to “go all the way” with the laxative toilet scene’: Jeff Daniels on Dumb and Dumber.

‘I was sitting next to my parents at a preview.

When we got to the toilet scene, my father put his head in his hands and said, “No, Jeffrey …”’

Jeff Daniels, played Harry

I was bored with drama and wanted to explore comedy, but my agents didn’t want me to do the film. They said: “There’s a chance Jim Carrey will wipe you off the screen.” I said: “Maybe – but not if I work with him.” And that’s what I did. I remember thinking: “It’s either going to be such a bad career move I’ll never work again, or it just might be funny.”

     One day Jim came into rehearsals having put a bowl on his head and cut his hair, because that’s how his character Lloyd would do it. I thought: “I gotta come up with something.” We were staying at a Holiday Inn, and I started using the shampoo without conditioner. Then I’d towel my hair dry and not comb it. And that was it. I came in and said, “Maybe this is Harry’s look?” Everyone agreed. After a couple of months, you could almost break my hair off.

      The first morning of shooting had me and Jim up against each other on a little scooter going down mountains. We had helmets on but I had no idea whether Jim could even drive a scooter – and then he started weaving!

     The scene where I got my tongue stuck to the icy chairlift was bumpy, too. It’s one of the few times where you can see me break into laughter because I just knew it was funny.

     I was stretching myself comedically. I remember getting ready to shoot the toilet scene, after my character unknowingly drinks laxative-laden tea. Jim asked me: “Hey man, you OK?” I said: “Just cold feet. This is either the beginning of my career or the end of it.” He said: “You’ve gotta go all in, man. Go all the way!” He couldn’t have been a better cheerleader. Me on that toilet will be the image on my tombstone.

     I have a theatre company in Michigan so I put on a preview. I was sitting next to my parents, and when we got to the toilet scene, my father hung his head in his hands and said: “No, Jeffrey …” Meanwhile, 5,000 people fell out of their chairs laughing. The reviews were horrible though. I still have a scrapbook of 200 newspapers panning the movie and wishing it never existed. Then we were the box office No 1 for six straight weeks. That’s when it hit me that we’d done the impossible.

Text (B)

Can NATO survive without the United States?

Europe is staring down the barrel of a stark new reality where the United States being the backbone of NATO – the alliance that has guaranteed the continent’s security for almost 80 years – is no longer a given.

President Donald Trump’s public animosity towards Ukraine’s Volodymyr Zelensky, his willingness to embrace Russia’s Vladimir Putin and recent comments casting doubt over whether he would defend NATO allies “if they don’t pay” have all forced European leaders to start thinking the previously unthinkable - is the US a reliable security partner at a time when the continent is being rocked by its biggest war since the 1940s?

     But NATO without the US is far from impotent, with more than a million troops and modern weaponry at its disposal from the 31 other countries in the alliance. It also has the wealth and technological know-how to defend itself without the US, analysts say.

The US and Germany are the biggest contributors to NATO’s military budget, civil budget, and security investment program, at almost 16% each, followed by the UK at 11% and France at 10%, a NATO fact sheet says. Analysts say it wouldn’t take much for Europe to make up for the loss of Washington’s contribution.

     If European countries unite and buy the right equipment, Europe “could pose a serious conventional and … nuclear deterrent” to Russia, Ben Schreer, Europe executive director of the International Institute for Strategic Studies (IISS), said in a Zoom call with CNN and other journalists in late February. “Europe alone (still has) a capacity to muster the resources it would need to defend itself, it’s just a question of whether (it is) willing to,” Schreer said.

And that’s the key question. Over more than 75 years and the administrations of 14 different US presidents, including the first Trump administration, the US has been the sinew that has kept the alliance together.

     During the Cold War, US troops on the continent were there as a deterrent to any Soviet ambitions to expand the Warsaw Pact alliance and eventually saw out its end when the Berlin Wall fell in 1989. NATO campaigns in the Balkans in the 1990s were conducted with US troops and airpower. And, until the second Trump administration took office on January 20, Washington spearheaded aid for Ukraine. Those decades of trans-Atlantic solidarity may have come to an end in recent days, analysts say.

Text (C)

Why old books are the most refreshing reads

When I was in 7th grade, I lied and said I liked “Pride and Prejudice.”

     I was secretly unimpressed, but I felt enough social pressure to appreciate perennially beloved classics to convince myself that I was smart enough to “get” the novel (I wasn’t). For a long time afterward, I kept a mental rant stored in my back pocket about how classics are boring, stuffy and way too long.

But in the past year or two, the strangest thing has happened to me: I found myself reading and — against all odds — actually enjoying books written in the 1800s. I’m honestly not sure what flipped the switch for me: better reading comprehension, more brain cells or accidentally becoming a pretentious, pseudo-intellectual college kid? I can’t be sure of the exact cause, but I suddenly felt way more inclined to understand life instead of just living it — and I think reading had a lot to do with it.

The books you consume are like thin layers of watercolor paint that gently seep into the canvas of your mind and color everything with a subtle tint. They change the way you look at life. In a world of muddled browns and tactile reds, where we judge a person by their political leanings, there’s always someone dying somewhere in the world and everyone is anxious to not offend anyone else — a headache by any standard — classics are like a puff of light blue smoke. They are the lingering remnants of an old world that somehow seems quieter, simpler and a bit more grounded than our current, overstimulated existence.

     There is something undeniably refreshing about these literary time capsules. It’s a pure, unedited glimpse into an older generation, people who were fundamentally just like us, albeit in a completely different context. It’s easy to relegate people who lived a long time ago into a totally different category — like when you wave away your grandma’s unabashed racism with an awkward laugh and a “she’s from a different era.” We so easily accept that so much of human belief and behavior is a result of the environment we were born into, so it’s not that hard to imagine that some future generation will stroll down the block and condemn everything we ever stood for as backward, ridiculous and possibly even pure evil. Some people might take this as proof that nothing matters and conclude that all our attempts at any real meaning are ultimately futile, but I just don’t think that is true.

     For all our generational and cultural differences, certain values and struggles still ring true. They strike a resonant chord that echoes into the present, even if our current social consciousness claims to have moved on from such archaic ideas. When Jane Eyre suppresses her passion to preserve her integrity, we can’t help but marvel at her strength of character — even as we live in a culture that would probably criticize her for “caving” to the oppressive social constructs of her day. When Mr. Knightley calls Emma out for being selfish and she actually hears it and apologizes, we intuitively know it’s beautiful to have a relationship that can withstand truth and confrontation — even if we ourselves tend to give up on our friendships after the first bout of relational conflict. When Oscar Wilde writes, “Women defend themselves by attacking, just as they attack by sudden and strange surrenders.” I can’t help but laugh and know exactly what he’s talking about — though I suppose it isn’t politically correct to do so.

What we often dismiss as obsolete is still very much human, and though the societal definitions of “good” and “bad” shift every couple of decades, we have so much to learn from the people who came before us. What did they make of the mystery that we call “life?” How did they understand it, and what about the human experience is universal, transcending time and circumstance? These are the wonderings that fill my brain when I read classics.

     A funny result of it all is that I feel more connected to humanity as a whole, all 117 billion of us who have ever walked and lived and breathed on this earth. We all want to love and be loved, to find purpose and meaning, to wrestle against the uglier side of ourselves and emerge victorious, someway, somehow. We want so many things, and we’ll do so much and travel so far to find what we’re looking for. It’s a breath of fresh, crisp air to read of the same human endeavor, but in such a different font -a different way of thinking. Time passes, and it’s funny and beautiful and exciting when old things have a chance to be wonderfully new again.