Exploring the Usability of Handwritten Arabic CAPTCHAs as Tools for Supporting Language Learning

Completely Automated Public Turing tests to Tell Computers and Humans Apart (CAPTCHAs) are among the most widely used applications of Artificial Intelligence (AI) for enhancing web security. Recently, Arabic CAPTCHAs have begun to attract research attention, though studies focusing on handwritten Arabic CAPTCHAs remain limited. This paper explores the usability of a proposed handwritten Arabic CAPTCHA model designed not only to differentiate between humans and bots but also to serve as a tool for supporting Arabic language learning. The model incorporates handwritten Arabic letters and words in a gamified format to aid in character recognition and engagement. To address the lack of usability evaluations in this area, the study empirically assesses the model's effectiveness and efficiency. Results indicate that meaningful handwritten text led to strong usability outcomes, with an average completion time of 4.7 seconds and an accuracy rate of 98.6%. Based on these findings, an abstract model of a language-learning-oriented CAPTCHA system is proposed. Additionally, the study identifies several usability challenges that should be addressed in future development to improve user experience and educational value