Digital transformation and advancements in Artificial Intelligence (AI) call for Arabic language assessments that measure not only linguistic proficiency but also communication skills, creativity, critical thinking, and digital literacy. However, Maharah Kalam assessments still tend to rely on memorized dialogues, question-and-answer sessions, and conventional presentations. This study aims to: (1) to analyze the implementation of STEAM-based authentic Maharah Kalam assessment through an AI-assisted comic project; and (2) to analyze the project’s contribution to the development of future-oriented learning competencies. This study employed a qualitative approach using a case study design involving 20 eighth-grade students in the International Class Program at MTs 2 Nurul Islam Mojokerto. Data were collected through observation, interviews, and documentation, and then analyzed using the Miles, Huberman, and Saldaña model. The results of the study show that an authentic STEAM-based Maharah Kalam assessment using an AI-assisted comic book project is capable of providing a contextual and meaningful evaluation of Arabic language proficiency. A total of 14 out of 20 students (70%) achieved a score of ≥70, with an average of 74.75. This project also develops future-oriented learning competencies, particularly creativity (80%), digital literacy (75%), communication (70%), confidence (65%), critical thinking (60%), and problem-solving (55%). Thus, the integration of STEAM and AI has the potential to serve as an innovative and future-oriented alternative for Arabic language assessment. This study contributes by proposing an authentic assessment model that integrates Arabic, STEAM, and AI, with AI serving as a cognitive tool and the teacher acting as a facilitator.