Reflective writing is a core pedagogical practice in higher education, fostering critical thinking, metacognition, and self-awareness across diverse disciplines. The emergence of generative artificial intelligence (genAI), particularly large language models, presents new benefits and risks for supporting reflective writing. This scoping review mapped the literature on genAI-facilitated reflective writing in higher education, examining its uses, opportunities and challenges, as well as guidance for educators. A systematic search including ERIC, PsycINFO, PubMed, Scopus, Google Scholar and key journals identified 39 studies published since 2021, predominantly from Asia. Analysis revealed five functional roles for genAI: eliciting, structuring, deepening, synthesising, and judging reflection. Opportunities include personalised prompts, scalable feedback, supportive learning dialogues, and potential efficiencies for educators, while challenges encompass ethical concerns including data privacy, biases, and potential for students' overreliance on genAI. The review highlights the central role of educators in guiding genAI use, preparing students, designing contextually appropriate tools, fostering metacognition, and maintaining ongoing oversight. Findings underscore that genAI can augment, but not replace, human-centred reflective practices. Future research should develop a more granular understanding of how genAI shapes the cognitive processes underpinning reflective writing. Overall, genAI offers transformative potential for reflective writing in higher education, but its deployment must be deliberate, pedagogically grounded, and ethically informed to maximise benefits while mitigating risks.