A Study of the Impact of ChatGPT on Translation Teaching

Abstract： ChatGPT is a new type of natural language processing machine learning model with powerful content production capacity. ChatGPT has the potential to change the traditional translation teaching mode， which can effectively assist the development of personalized teaching， promote students' independent learning， build a new type of teaching subject relationship， innovate the teaching management mode， promote the change of the mode of knowledge production， and broaden the space of translation education and teaching research. At the same time， there are potential risks in the application of ChatGPT， such as weakening learning autonomy， alienating the relationship between teaching subjects， damaging academic integrity， and triggering risks of knowledge security， ethics and ideology. In order to cope with the challenges and give full play to the potential of ChatGPT in translation teaching， it is necessary to innovate the teaching mode， reform the teaching process， enhance students' independent learning ability， strengthen the education on translation ethics and ideological risks， and improve the digital literacy of teachers and students. The new artificial intelligence represented by ChatGPT as a "human-like teacher" has powerful teaching potential， but it cannot replace the human factor in education and teaching， and thus cannot change the essence of education， and teachers will still play an irreplaceable role in many aspects in the future.

Keywords：ChatGPT; generative artificial intelligence; translation teaching; digital literacy; educational technolog

1. Introduction：

ChatGPT has the ability to integrate and synthesize material， develop integrated perspectives， and engage in an ongoing dialog with humans based on the integration of large amounts of data and training applications. These capabilities give ChatGPT a wide range of educational applications. It has been shown that ChatGPT can effectively improve students' language skills （Kohnke a ， 2023 Jiashen 2023：4） pointed out that ChatGPT has upgraded from the primary stage language text to the advanced stage of logical， semantic and shell to the selection and judgment of language content f knowledge processing that initially possesses the natu foreign language education. Translation teaching chin an important part of foreign language educatio challen s brought by ChatGPT to translation teaching d the relat colle in the academic field. 2. ChatGPT and Translation Teaching Change

As a new type of artificial intelligence， ChatGPT has a wide range of application prospects in translation teaching and has great potential in promoting the personalized development of teaching， reconstructing the relationship between teaching subjects， changing the concept of teaching and expanding the space for research in translation education and teaching.

2.1 Promoting the personalization of teaching and learning

ChatGPT can help build individualized learning environment （Zhang Zhenyu and Hong Huaqing， 2023：39） and promote the personalized development of teaching and learning （Feng Zhiwei and Zhang Dengke， 2023：6）.ChatGPT has a strong dialogue ability， and shows a good human-computer interaction performance. Teachers can use ChatGPT to deeply analyze the traces of students' human-computer interaction in the process of translation learning， comprehensively analyze the content direction， focus， discourse characteristics and frequency pattern of students' questions， and judge the students' knowledge structure， learning styles， motivation and cognitive characteristics， etc.， so that they can recommend learning resources， provide guidance on learning methods， and guide the paths and directions of professional development for the students in a targeted way. Based on the behavioral data of students' ChatGPT learning， teachers can gain an in-depth understanding of students' learning status， dynamically track the evolution of students' translation level and the characteristics of changes in their translation behaviors， and accurately portray the translation ability of different students， so as to lay the foundation for maximally stimulating students' learning potential. It has been shown that artificial intelligence chatbots can assist students in setting learning goals （Hew， et al.， 2023） and assessing their second language vocabulary levels （Jeon J， 2021）， which is conducive to independent learning. Students with the awareness and ability of independent learning can independently complete learning tasks to achieve their learning goals， and show a high degree of self-exploration， self-selection， self-construction， and self-creativity.ChatGPT's excellent intelligent response ability， multi-round dialogue ability， and intelligent content generation ability can effectively promote students' independent learning and strengthen students' learning autonomy. An important goal of translation teaching is to develop students' petence. Translator competence consists of linguistic， textual， disciplinary， cultural， （Neubert， 2000 ， instrumental， strategic and cognitive competencies （PACTE Group， 2005： 612）. s and development goals of their own translator competence， students development of translator competence， make a study schedule， plan the content study， build a professional knowledge system and test their translation practice ability with the support of ChatGP o as to improve their professionalism in translation.

2.2 Reconstructing the Relationship between Teaching Subjects

On the one hand， ChatGPT has given birth to a new type of teaching subject. The traditional AI teaching machine adopts the working mode of searching， retrieving and sequencing to respond to users' needs， which lacks flexibility， mobility and interactivity. In contrast， ChatGPT has more prominent information processing capability， semantic parsing capability， content generation capability and sequence task execution capability. Sequential task execution capability means that ChatGPT is able to understand the correlation between instructions according to the user's description of sequential instructions， advance the task step by step， and complete complex tasks based on combination of instructions， so that it can better execute multi-step sequential tasks given by the user （Lu et al.， 2023：27）. These capabilities enable ChatGPT to understand the user's intention and generate meaningful text based on contextual information around a given topic in a multi-round dialog with the user. This also allows C hatGP to participate extensively in the translation teaching process as an intelligent teaching subject. Before class， ChatG can assist teachers in preparing lessons， compiling syllabus and lesson plan， setting teaching contents adjustin ourse knowledge structure， enriching translation teaching cases， and simulatin eachin activit duri T can assist teachers in explaining translation cases， analyzing translation analyzin ting pe sonalized translation tasks for students， evaluating student highl nti ting problems， and providing students with personalized translation teachers in explaining translation cases， analyzing translation sty behavi ettin per onalized translation tasks for students， evaluating students' translation problems， and solving the linguistic and intellectual difficulties enc ter udents in the roces slatio n; after the lesson， ChatGPT can assist teachers in tracking students' learning status， evaluating the teaching effect， and providing students with academic counseling.

On the other hand， ChatGPT constructs a new type of human-machine relationship that is interactive and inter-constructive. In traditional intelligent teaching systems， AI teaching robots do not have human-like mobility and cannot be included in human contexts， and human-machine interaction is mainly manifested as a one-way programmed response of the machine to the commands issued by human beings. ChatGPT， on the basis of super-large scale pre-trained corpus and through the deep learning mechanism， has gained powerful semantic analysis capability， and has realized the function of entering human context and understanding human language. In terms of operation logic， ChatGPT applies the training mechanism of "reinforcement learning based on human feedback"， and it can acquire human language characteristics and way of thinking in the process of interacting with human beings， so as to evolve its ability. From the perspective of learning， the process of human-computer interaction in teaching and learning activities is a process of mutual learning and construction between the two. In teaching， teachers and students can use ChatGPT to acquire translation corpus， analyze translation style， correct translation errors， and solve the difficulties in professional knowledge in translation， so as to improve their knowledge， thinking and ability， etc. ChatGPT acquires training corpus in the process of interacting with students and teachers， and evolves in the aspects of information capacity， corpus structure and inventory， and semantic analysis ability through the machine learning mechanism. ChatGPT shows a certain degree of subjectivity and initiative in the process of human-computer interaction. This new human-machine relationship requires us to position the identity of the AI teaching robot represented by ChatGPT with new concepts， and we should regard ChatGPT as an intelligent teaching subject that grows together with us， not just a mechanical teaching tool.

2.3 Expanding the space of translation education and teaching research

ChatGPT has been widely used in translation teaching， changing the traditional dual teaching mode of teacher-student and making the translation classroom move towards the teaching mode of "teacher-machine-student" interaction and synergy， and the construction of human-computer teaching community puts forward a new topic for research on translation education and teaching. The construction of the human-machine teaching community has raised new issues for the study of translation education and teaching. how should we deal with the relationship between ChatGPT， as a human-like teaching subject， and the human teaching subject？ How is its subjectivity defined， realized， measured and affected？ Other questions that arise from these issues of subjectivity include： Does ChatGPT， as an AI tutor for translation teaching， have the power to independently evaluate students' translation assignments and test papers？ How does ChatGPT determine the independence and workload of students when participating in their translation practice or translation research？ Are machine translation assignments considered to be the academic output of the student's studies？ How are such outcomes evaluated from a professional perspective？ Do the teaching resources generated by ChatGPT （e.g. translation cases and reference translations， etc.） have intellectual property rights？ Who bears the ethical， moral and political risks of ChatGPT-generated translation teaching resources？ In addition， while previous studies focused on the "interpersonal" relationship in education and teaching， the "human-computer" relationship in the new teaching and learning scenario also contains a wealth of research topics， such as the cognitive mechanism of human-computer interaction， the discourse characteristics of human-computer dialogues and the underlying logic of human-computer emotional interactions. logic of human-computer emotional interaction， etc. Researchers need to conduct a comprehensive study on the translation teaching ecosystem centered on "teacher-machine-student"， and it is also necessary to explore the relationship between the cognitive， psychological， emotional and behavioral characteristics of teachers and students and the application scenarios of ChatGPT， and to increase the research on the application scenarios of translation teaching with the new artificial intelligence technology.

3 Path to Cope with the Challenges

In order to cope with the series of risks that may be caused by ChatGPT， it is necessary to build a harmonious teacher-student relationship， form an ethical risk management mechanism， improve the ability to maintain knowledge security， and establish an ideological risk prevention and control guardrail.

3.1 Strengthen emotional interaction and build a harmonious teacher-student relationship

Teaching is an emotional interaction process between teachers and students， and the emotional significance and humanistic value generated in the process of thought provocation， collision of ideas， fusion of feelings， intercommunication of hearts and mutual shaping of personalities between teachers and students is an important content in addition to the transmission of knowledge. ChatGPT does not have self-consciousness， subjective intention and moral emotions， and although it is able to generate emotion-like information inspired by the students' questions， it cannot resonate with students' thoughts， feelings and psychological rhythms in the light of the specific teaching environment. Although ChatGPT can generate emotion-like messages inspired by students' questions， it cannot resonate with students' thinking， emotions and psychological rhythms according th teaching and learning environments. the emotion-like messages generated by ChatGPT duri its algorithmic responses to human emotions by means of simulation technolog which odels and program codes. ChatGPT lacks true language comprehension and cannot ndersta f human emotions embedded in language.

Funding

This work was supported by Artificial Intelligence Translation Shaanxi University Engineering Research Center Reference

［1］HEW K F ，HUANG W ，DU J ，et al. Using chatbots to support student goal setting and social presence in fullyonline activities：Learner engagement and perceptions［J］. Journal of Computing in Higher Education，2023 （1）：40-68. ［2］Jeon J. Chatbot-assisted dynamic assessment（CA-DA ）for L2 vocabulary learning and diagnosis［J/OL］. Foreign Language World No.2，2023（ General Serial No.215 ） Computer Assisted Language Learning，2021 ［2023-03-12 ］. https：//doi.org/10.1080/ 09588221.2021.1987272. ［3］KOHNKE L ，MOORHOUSE B L & ZOU D. ChatGPT for language teaching and learning［J/OL ］. RELC Journal ，2023 （ 2 ）： 1-14 ［ 2023-06-06 ］ . https：//www. Researchgate. net/publication/ 369743021_ ChatGPT_ for_Language_Teaching_and_ Learning. ［4］MIJWIL M M，HIRAN K K，DOSHI R ，et al. ChatGPT and the future of academic integrity in the artificial intelligence era：A new frontier［J/OL ］. Al-Salam Journal for Engineering and Technology ，2023（2 ）： 116-127 ［2023-04-25 ］. https：//www. researchgate. net/publication/369977556_ ChatGPT_and_ the_Future_ of_ Academic_ Integrity_in_the_Artificial_Intelligence_Era_A_New_Frontier. ［5］NEUBERT A. Competence in language ，in languages ，and in translation［C］// SCHÄFFNER C & ADAB B. Developing Translation Competence. Amsterdam：John Benjamins，2000 ：3-18.