This study explores the application of AI-assisted metacognitive strategies by postgraduate students using GenAI tools in academic reading, with a focus on both taxonomy and associated emotional responses. Fifteen postgraduate students participated in an experiment that included pre-training, selecting and reading academic papers, interacting with a GenAI chatbot about academic reading, and concluding with post-experiment reflective interviews. Data collection methods included screen captures, AI prompts, and interviews, which were analyzed through thematic and sentiment analysis, as well as word cloud visualization. The identified taxonomy of AI-assisted metacognitive strategies in academic reading comprises: (1) AI-assisted planning strategy, involving selective attention and goal setting; (2) AI-assisted monitoring strategy, ensuring AI responses align with reading objectives and adjusting behaviours for enhanced comprehension; (3) AI-assisted evaluation strategy, where students critically assess both AI suggestions and their understanding; (4) AI-assisted support strategy, incorporating techniques such as note-taking and summarizing to interpret and manage AI outputs; and (5) AI prompting strategy, which activates a metacognitive cycle leveraging all previous strategies to optimize AI prompts. Additionally, the results indicated that AI-assisted planning, monitoring, and evaluation strategies were linked to positive emotions due to an effective feedback loop. However, AI-assisted support and AI prompting strategies elicited less positive emotions, primarily due to challenges in AI literacy and the cognitive demands of managing AI interactions. This research highlights the potential of GenAI tools in enhancing academic reading by facilitating the use of AI-assisted metacognitive strategies while underscoring the need for improved AI literacy and integrated pedagogical frameworks.