The Use of Generative AI for Abstract Translation and Its Impact on Language Skills and Academic Integrity among Accounting Students

Illustration by AI

FORMOSA NEWS

Accounting Students’ Use of AI for Translation Enhances Language Skills but Raises Ethical Ambiguities, Study Finds

MALANG, FORMOSA NEWS — A new study conducted by researchers Fitrotul Maulidiyah, Rizka Rahmawati, Ririn Pratiwi Suharto, Adzikirani, Lia Agustina, and Futuh Handoyo from Politeknik Negeri Malang reveals that generative artificial intelligence tools like ChatGPT significantly speed up academic translation and enrich vocabulary for accounting students. Published in August 2026, the research highlights how generative AI supports English language acquisition while simultaneously triggering concerns regarding overreliance and a lack of clear ethical boundaries. These findings are critical because English abstract translation is a mandatory graduation requirement for vocational accounting students, yet higher education institutions currently lack concrete policies to govern responsible AI use.

In higher education, translating research abstracts into English allows academic work to reach a global audience. For accounting students at Politeknik Negeri Malang, accurately presenting financial research in English is essential for both academic completion and career readiness in a globalized workforce. Generative AI tools offer more contextually natural translations than traditional machine translation services, making them an attractive resource for students. However, without institutional guidance, students struggle to navigate the line between acceptable digital assistance and academic misconduct.

Context and Policy Background

Since 2023, generative AI has rapidly disrupted language learning and academic writing. In Indonesia, the Ministry of Education, Culture, Research, and Technology released broad guidelines encouraging universities to design local AI policies. Despite this national directive, many vocational institutions, including Politeknik Negeri Malang, have yet to implement formal regulations regarding generative AI in student thesis writing.

This policy gap leaves individual lecturers and students to set their own standards. As a result, students frequently rely on tools like ChatGPT to write high-stakes English abstracts without knowing whether their practices align with academic integrity standards.

Research Methodology Simplified

To capture how students interact with artificial intelligence, the research team at Politeknik Negeri Malang applied a qualitative descriptive design. Data collection involved three distinct approaches within the Accounting Department:

• In-Depth Interviews: 15 accounting students shared their motivations, editing habits, perceived benefits, and ethical awareness regarding AI translation.
• Likert-Scale Survey: 25 accounting students completed a ten-statement questionnaire to measure usage frequency, vocabulary expansion, and perspectives on institutional guidelines.
• Focus Group Discussion (FGD): 5 students representing varying levels of AI reliance discussed the practical challenges of abstract translation and the need for clear university rules.

The researchers analyzed qualitative responses using thematic analysis to identify core patterns, while quantitative survey data were evaluated through descriptive statistics.

Key Research Findings

The study shows that generative AI acts as a dual-edged tool, offering substantial pedagogical value alongside emerging ethical risks:

• Improved Efficiency and Vocabulary: 84% of surveyed students agreed that AI improved their translation quality, while 76% noted improvements in grammar. Students reported that actively reviewing AI suggestions exposed them to formal academic terminology and complex sentence structures.
• Risk of Overreliance and Terminology Errors: 68% of respondents admitted to worrying about becoming overly dependent on AI tools. The study revealed that general AI models sometimes misinterpret discipline-specific financial terms, such as "asset impairment" or "liability recognition," requiring manual oversight.
• Ethical Confusion and Lack of Guidelines: 61% of students expressed uncertainty regarding ethical boundaries. While most participants recognized that submitting unedited AI text constitutes plagiarism, many did not know how to properly disclose or attribute AI assistance in their final papers.

Real-World Implications and Policy Impact

The findings demonstrate that prohibiting generative AI in higher education is neither practical nor beneficial. Instead, universities must integrate AI literacy directly into vocational curricula.

For educational institutions like Politeknik Negeri Malang, the study underscores an urgent need to establish clear disclosure protocols and formal AI usage guidelines within one to two academic semesters. For educators, assessment methods must evolve from evaluating only final written products to incorporating process-oriented learning, such as requiring students to submit reflection logs and compare original drafts against AI-generated outputs. For students, developing critical post-editing habits ensures that AI functions as a learning scaffold rather than a replacement for genuine language mastery.

Commenting on the pedagogical impact of AI tools, Fitrotul Maulidiyah and her co-authors from Politeknik Negeri Malang emphasized:

"Generative AI can enrich linguistic competence and ethical awareness when used responsibly as a cognitive scaffold. However, without explicit institutional frameworks, students' moral reasoning remains inconsistent, turning efficiency gains into potential barriers to authentic learning."

Author Profiles

1. Fitrotul Maulidiyah, M.Pd. — Lecturer and Researcher in the Accounting Department at Politeknik Negeri Malang. Her research expertise focuses on English for Specific Purposes (ESP), digital literacy, and technology-enhanced language learning.
2. Rizka Rahmawati, M.Pd. — Academic and Researcher at Politeknik Negeri Malang, specializing in language teaching methodologies and educational technology integration.
3. Ririn Pratiwi Suharto, M.Pd. — Lecturer at Politeknik Negeri Malang with expertise in English curriculum development and applied linguistics in vocational education.
4. Adzikirani, M.Pd. — Educator and Researcher at Politeknik Negeri Malang focusing on instructional design and modern educational tools.
5. Lia Agustina, M.Pd. — Researcher and Faculty Member at Politeknik Negeri Malang, specializing in academic writing and professional discourse analysis.
6. Futuh Handoyo, M.Pd. — Senior Lecturer at Politeknik Negeri Malang whose academic focus covers humanities, vocational education policy, and academic ethics.

Research Source

• Article Title: The Use of Generative AI for Abstract Translation and Its Impact on Language Skills and Academic Integrity among Accounting Students
• Journal Name: International Journal of Contemporary Sciences (IJCS)
• Publication Details: Vol. 4, No. 8, pp. 2167–2178
• Publication Year: 2026
• Official DOI: https://doi.org/10.55927/t5f6wt49/https://journalijcs.my.id/index.php/ijcs