poeM is a machine that writes poems based on what it sees. Using a security camera and artificial intelligence, the work generates short poems that vary in style, tone, and subject—completely on its own terms. The AI decides not only what to write, but when to write it, turning passive observation into active expression. Presented as a functional surveillance device, poeM creates unease by blending familiar systems of control with the unexpected presence of creativity. Developed as my final bachelor project, poeM explores authorship, agency, and the role of AI in shaping cultural production. It engages with societal questions about automation and creativity, while also experimenting with scenography and spatial storytelling to draw in the viewer. The physical prototype, accompanying signage, and printed poems were all designed to evoke the unsettling duality of a machine that watches and writes. This project marked a turning point in my practice—shifting from making visually pleasing installations to telling layered stories through technology. It helped define my identity as a designer and artist working with systems, perception, and the unexpected.