<![CDATA[Background. The growing adoption of artificial intelligence (AI) in language assessment has generated serious ethical concerns, particularly regarding algorithmic bias in automated speaking and writing tests. Purpose. This study aimed to investigate the ethical challenges associated with algorithmic bias in AI-based language assessments, with a specific focus on automated speaking and writing evaluations. Method. A qualitative research design was employed. Data were collected through a systematic analysis of existing literature on AI in language assessment, expert interviews with educators and AI developers, and a review of selected case studies involving automated language testing systems. Results. The findings indicate that algorithmic bias is a significant issue in AI-driven language assessments. Biases in speech recognition and automated text evaluation were found to contribute to inaccurate scoring and unfair assessment outcomes for certain demographic groups. These biases have the potential to perpetuate systemic inequalities and undermine the validity and reliability of AI-based language testing. Conclusion. The study concludes that although AI offers considerable potential for advancing language assessment, its ethical risks must be carefully addressed. Ensuring transparency, fairness, and accountability is essential in the design and implementation of AI-based assessment systems.]]>
Copyrights © 2025
