This study explores the experiences of tertiary students in Benguet in using Artificial Intelligence (AI). Employing a mixed-method research design, the study gathered quantitative data through electronic surveys from 66 tertiary students and qualitative data via semi-structured interviews with 12 students from three Higher Education Institutions in Benguet. The research analyzed the types of AI tools used, how students utilize them, the challenges encountered, and their adaptive strategies.
The findings reveal that students commonly use generative AI tools such as ChatGPT, Cici, and Grammarly primarily for information acquisition, grammar checking, and enhancing outputs. Efficiency and convenience is identified as a common reason why students use these AI tools. This study higlights two main approaches to AI including AI-assisted shortcutting, where students heavily rely on AI-generated output with minimal intervention, and Human-in-the-loop (HITL), involving more human control through collaborative ideation, content improvement, text mining, and plagiarism detection. The study highlights several challenges, including limited accessibility to premium AI features, and inconsistent internet connectivity. It identified perceived challenges such as inaccuracies and insufficient data on specific topics like indigenous knowledge, lexical complexity of AI language, and issues such as the potential for AI dependency and plagiarism. In response to these challenges, students employ cross-referencing with traditional resources and tool-chain integration to verify information and enhance AI-generated content.
The study concludes that while generative AI offers convenience, its misuse can lead to inaccurate outputs, AI dependency, and plagiarism. It recommends that educational stakeholders should develop AI policies and implement AI literacy programs to guide students toward the effective and ethical integration of AI in their learning.