Poemxreader Andrea Fernandez Preview Set against various considerations of the poetic and the digital, poemxreader is a Twine-based interactive piece that reflects on the interplays between reading and writing, cognition and information processing. The project surrounds the poetic capacities of ChatGPT as an NLP (Natural Language Processing) tool, exploring the ways in which its mechanisms may resonate with or rub against longstanding poetic conventions in “unconscious” or free associative writing practices. To this end, it looks into the processes by which Poem Writer 5.1. “predicts” the next most likely word in a given response sequence, positioning this as a potential analogue or counterpoint to cognitive free association. Here, the prompt becomes a stable frame, moving across controlled variations in voice, rhyme, theme and metaphor, and opening up later into free associative reflection. Taking viewers through a series of “conversations” between a semi-fictional ‘User’ and the AI ‘Poem Writer’, the passages in poemxreader are constructed along the usual features of generative writing at large; all movements are dependent on the reader’s choices. Mingling the formal features of poetry, prompt-and-response and gameplay/interactive text, this opens up insights into the dynamics of reading and writing as formed in relation to each other, urging us to question their conventional subject-positions through the synthesis of AI-generated poetry Author Profile Andrea is an undergraduate student at Ashoka University, pursuing English, Creative Writing, Performing Arts and Political Science. GO TO WORK