Mobile random text-based voice authentication for older adults: A pilot study

Abstract

Thai older adults often struggle to use existing smartphone authentication systems due to age-related problems. Random text-based voice authentication, which resists Replay attacks and reduces the need to memorize information, is a suitable alternative. Nevertheless, past studies have insufficiently focused on usability, leading this research effort to improve the usability of voice authentication for older adults. As part of the research, this study seeks to develop a random text-based voice authentication suitable for older adults. The prototype was implemented as a mobile application, VAuth, with Voice User Interface (VUI) as a primary interaction method. Formative Usability Testing was conducted iteratively and incrementally to evaluate and refine VAuth. The first version of VAuth was well-received despite initial problems, which were resolved subsequently. Overall results indicated that VAuth is usable for older adults. Participants could complete all given tasks (enrollment and verification) through VUIs. They responded favorably to VUIs in general and acknowledged their advantages. Moreover, the results from the SEQ and SUS questionnaires aligned with participants' positive opinions. The dialogue design was adequate, and older adults could correctly pronounce Thai-word passphrases in most cases. While the accuracy of Thai speech recognition remains an issue, progress in voice recognition may eventually overcome it. Despite these encouraging results, this study still has limitations. While the test helped determine if VAuth is a viable solution, it did not confirm whether VAuth remains usable when deployed in larger, different contexts, such as a noisy environment or public places. Thus, further studies are needed to compare VAuth with existing authentication methods with an enlarged sample size in multiple usage scenarios.