Pemilihan Artificial Intelligence Tools Menggunakan Metode SAW Berbasis Web Pada Mahasiswa UNBARA

Engineering notes

Engineering notes will be added by the aipentium editorial team.

Chinese explanation / 中文解读

中文解读待补充：本站会优先为大语言模型、生成式AI、ChatGPT相关技术、计算机视觉、深度学习等高价值论文补充中文说明。

Original abstract

The rapid development of artificial intelligence (AI) tools presents a challenge for students in selecting the tools best suited to their academic needs. This study aims to recommend the best AI tools for students at Universitas Baturaja using the Simple Additive Weighting (SAW) method. Data were collected through questionnaires distributed to 150 students at Universitas Baturaja, selected using a purposive sampling technique, namely students who had used at least two of the four evaluated AI tools in their academic activities over a certain period of time. The assessment was carried out on four alternatives Meta AI (A3), and Claude AI (A4), based on five criteria: ease of use (K1, weight 0.20), response speed (K2, weight 0.20), feature completeness (K3, weight 0.20), answer accuracy (K4, weight 0.25), and cost (K5, weight 0.15). Criteria weights were determined through expert discussion, taking into account the relative importance of each criterion to students' academic needs, while the score of each alternative on each criterion was obtained from the average rating given by 150 respondents on a 1–5 Likert scale. The SAW calculation results show that ChatGPT obtained the highest preference value of 0.9519, followed by Google Gemini at 0.8774, Claude AI at 0.8533, and Meta AI at 0.7319. ChatGPT performed best because it received the highest scores on the feature completeness and answer accuracy criteria, and achieved a perfect normalized value (1.000) on the cost criterion, since ChatGPT had the lowest cost score among the alternatives, making it the minimum reference value in the cost normalization calculation; meanwhile, Google Gemini performed better on ease of use and response speed. The SAW method in this study was able to produce a ranking that is structured, consistent, and mathematically traceable based on user perception, making it suitable as one of the institutional references for supporting AI tool selection, with the caveat that the results still reflect the subjective perceptions of respondents during the particular data collection period.

Links and sources

Comments