教育機器人在課堂中已被證實能有效提升學生學習動機與參與度。然而，傳統教育機器人仍面臨幾項挑戰：互動受限於實體空間與課堂時間；資源限制導致學生難以擁有專屬且個別化設定的機器人；互動設計多聚焦於短期任務，缺乏引導學生朝長期目標前進的機制，學習動機難以持久。本研究提出整合實體與數位形式的AI寵物機器人，並命名為Self-Agent for Growth & Expertise with Resonance(SAGE-R)。透過多管道學習機制，提供學生在課堂內外皆能持續互動的學習體驗，突破時間與空間限制。融合依附理論與心理擁有感，學生可個別化寵物外觀、聲音與個性，透過精靈「附身」概念確保虛實載體中角色一致性，系統整合ChatGPT與未來自我理論，以成為領域專家為未來目標，寵物作為現在與未來自我共鳴的橋梁，配合課程大綱與學習評量設計學習活動，學生為照顧寵物與未來自我而投入學習。根據社會相互依賴理論，安排與寵物共同展演的最終目標，使學生與寵物合作學習直到最後。本研究於桃園某科技大學針對101名餐旅管理系學生進行為期十八周的實驗，透過前後測、問卷與訪談進行多元評估，並以SPSS分析。結果顯示，相較於傳統AI學習夥伴，本SAGE-R系統更能增進學生的情感連結與責任感，引導學生將當下努力與長期目標連結，穩定參與學習並提升成效。;Educational robots can enhance students’ motivation and engagement. However, traditional designs face limitations due to classroom time and space constraints, resource constraints, and a lack of long-term goal orientation. This study proposes Self-Agent for Growth & Expertise with Resonance (SAGE-R), a personalized AI pet robot system. SAGE-R integrates physical and virtual interaction channels, enabling students to learn anytime and anywhere. Based on Attachment Theory and Psychological Ownership, students can personalize the genie’s appearance, voice, and personality. The system incorporates the Future Self Theory, guiding students toward a future expert identity. The pet creates a sense of resonance between students′ present and future selves. Learning activities align with the syllabus and rubric, encouraging students to care for both the pet and their future self. Based on Social Interdependence Theory, a collaborative final performance with the pet fosters long-term cooperation. A 18-week experiment with 101 hospitality management students in Taoyuan, using pre-/post-tests, questionnaires, interviews, and SPSS analysis, shows that SAGE-R effectively sustains students′ learning motivation and improves outcomes.
