https://www.journal.sncopublishing.com/index.php/brillojournal <p align="justify"> </p> <p align="justify"><strong>Brillo Journal</strong> committed to providing a streamlined submission process, rapid review and publication, and a high level of author service at every stage. This journal is published twice a year (June and December) by <strong>S&amp;CO Publishing </strong>(a company in the publishing industry under the business license of <strong>CV. Samuel Manurung and Co</strong>) in collaboration with the <strong>Indonesian Society of Researcher and Educator.</strong></p> S&Co Publishing en-US Brillo Journal 2809-8528 <p>The authors agree that this article remains permanently open access under the terms of the Creative Commons Attribution 4.0 International License</p> https://www.journal.sncopublishing.com/index.php/brillojournal/article/view/169 <p>The rapid adoption of artificial intelligence (AI) in higher education has transformed students' learning practices, particularly in completing academic assignments and projects. However, the extent to which AI use influences academic performance remains inconclusive. This study investigated the effects of overall AI usage, the percentage of AI use in course projects, and mathematical disposition on undergraduate students' Grade Point Average (GPA). A quantitative correlational design was employed involving 39 undergraduate students. Data were collected using a validated AI usage questionnaire, a mathematical disposition questionnaire, self-reported percentage of AI use in course projects, and students' GPA. Multiple linear regression analysis was conducted after verifying the assumptions of normality, multicollinearity, and independence of residuals. The regression model was statistically significant (F = 6.247, <em>p</em> = .002) and explained 34.9% of the variance in GPA (<em>R</em>² = .349). Among the predictors, only the percentage of AI use in course projects significantly predicted GPA (<em>β</em> = −0.614, <em>p</em> &lt; .001), whereas overall AI usage (<em>p</em> = .651) and mathematical disposition (<em>p</em> = .695) were not significant predictors. These findings indicate that AI use itself is not associated with lower academic performance. Instead, students' reliance on AI in completing course projects appears to be associated with GPA. Excessive dependence on AI-generated outputs without critical evaluation, revision, verification using credible academic sources, and integration of students' own reasoning may reduce the quality of academic work. Therefore, AI should be utilized as a learning assistant that supports critical thinking, creativity, and problem-solving rather than replacing students' independent intellectual engagement. The findings highlight the importance of promoting AI literacy and responsible AI use in higher education.</p> Santri Purba Copyright (c) 2026 Santri Purba https://creativecommons.org/licenses/by/4.0 2026-06-30 2026-06-30 5 2 27 43 10.56773/bj.v5i2.169 https://www.journal.sncopublishing.com/index.php/brillojournal/article/view/177 <p>This study aims to analyze elementary school teachers' perceptions of deep learning in mathematics. This qualitative study was conducted at SDN Cipinang Melayu 09 Pagi, Jakarta (May-December 2025). Fourteen elementary school teachers were selected using purposive sampling. Data were collected through questionnaires and in-depth interviews, with method triangulation employed to ensure validity. Data analysis followed qualitative research principles through reduction, categorization, presentation, and conclusion drawing. All respondents demonstrated sound understanding of deep learning as an approach focusing on mindful, meaningful, and joyful learning and acknowledged its potential benefits in enhancing concept understanding, learning motivation, and problem-solving abilities. Teacher readiness varied, with high readiness (85.7%-100%) in basic teaching skills and collaboration, and moderate readiness (78.6%-85.7%) in practical implementation. Major barriers included infrastructure limitations (71.4%), administrative burden (57.1%-64.3%), and insufficient training (78.6% had attended related training). All respondents (100%) expressed strong expectations for specialized training, adequate devices, government support, and user-friendly platforms. While teachers possessed positive perceptions and sound conceptual understanding of deep learning, effective implementation requires comprehensive systemic support including intensive training, adequate infrastructure, user-friendly platforms, policy support, reduced administrative burden, continuous mentoring, and communities of practice. With such support, deep learning possesses significant potential to transform elementary mathematics instruction.</p> Novi Andri Nurcahyono Sani Sahara Wily Wandari Copyright (c) 2026 Novi Andri Nurcahyono, Sani Sahara, Wily Wandari https://creativecommons.org/licenses/by/4.0 2026-06-30 2026-06-30 5 2 15 26 10.56773/bj.v5i2.177