Evolusi Multimodal AI dalam Pendidikan: Inovasi GPT-5, Gemini Ultra, dan Transformasi Pembelajaran Digital

The development of artificial intelligence (AI) technology continues to experience significant advancements, particularly with the emergence of multimodal models such as OpenAI’s GPT-5 and Google DeepMind’s Gemini Ultra. These innovations mark the beginning of a new phase in AI utilization, where the technology is not only capable of understanding and generating text, but also processing images, audio, and video simultaneously. This capability opens up numerous opportunities in the field of education, especially in creating more interactive, personalized, and adaptable learning models tailored to students’ needs. The purpose of this article is to examine how the evolution of multimodal AI impacts digital-based educational practices, with a special focus on the potential and applications of GPT-5 and Gemini Ultra in the teaching process. A literature review method is employed to analyze the latest scientific and technical publications regarding AI implementation in education. The analysis focuses on technological capabilities, types of learning interventions, and potential risks and challenges that may arise. The findings indicate that multimodal AI can not only enhance the accessibility and effectiveness of learning but also facilitate the creation of more inclusive and contextual learning experiences. However, the implementation of such technologies requires strong regulations and ethical norms to prevent issues such as data privacy violations, unequal access to technology, and reduced human interaction in educational settings.Therefore, this article recommends a balanced use of multimodal AI—one that harmonizes technological innovation with human-centered pedagogical principles—to ensure that the digital transformation in education proceeds sustainably and equitably.