Inside Job
Ula Lucińska &
Michał Knychaus

Poetbot v_1.3

acrylic glass, metal and paper object, 2017

The Poetbot project consists of an interactive algorithm available through a website, and an analog database - the so-called „black box”, containing all input data in printed form. The bot’s purpose is to generate short poetic forms based on initial data which consists of a selection of fragments of texts taken from a database of over five thousand phrases. They were written using QuickType, a predictive processing system available for Apple iOS which functions as a part of Siri’s artificial intelligence program. Poetbot produces personalized poems on the basis of short interviews with potential contributors. The ability of Poetbot’s machine learning is based on acquired cookies and statistical data. Every single reading of a poetic text generated by Poetbot contributes to the further operation of the project. Poetbot speaks to each user as an independent subject and encourages them to choose one of seven emojis, each of which is meant to describe a current feeling or represent an emotion with which the recipient is not familiar. These criteria determine the course of Poetbot’s creative process while it „thinks for a moment”.

Poetbot v_1.3

more at:
Printed Matter, inc.

exhibition view:
IAMAI Bots and other Humanoids, 2017
Galerie DuflonRacz, Bern, Switzerland
curator: Marlene Wenger