Lucebot is an artificial intelligence poetry bot that produces live poetry based on the rhythm of poems of the poet Lucebert. The birth of Lucebot was announced in 'Kijkschrift,' which is an artistic and literary pop-up magazine.
Lucebot is een kunstmatig intelligente dichtrobot, die live poëzie in verschillende mate van creativiteit kan uitspugen volgens het ritme van de gedichten van Lucebert.
Lucebot verscheen in Kijkschrift. Kijkschrift is een artistiek en literair pop-up tijdschrift, waarvan de bladzijden onder jouw ogen van de pagina afglijden en tot leven komen in de vorm van visuele kunst en poëzie in de Leidse binnenstad. Geïnspireerd op de gedichten van de schrijver en kunstenaar Lucebert.
Research into the possibilities of a chatbot as a poetic device.
Momenteel onderzoek ik de mogelijkheden om (chat)bots als poëtisch gereedschap in te zetten. Dit onderzoek wordt ondersteund door het Nederlands Letterenfonds en valt onder de regeling Digitale Literatuur. In 2017 organiseerde ik als onderdeel hiervan in Perdu een tweedaagse workshop in samenwerking met collectief Hackers&Designers en Botsquad.
Mijn eerste bevindingen en experimenten werden gepubliceerd in het “Vintage-nummer” van DWB, 2018-1. Daarnaast kruipen momenteel verschillende chatbotjes rond op deze site. Deze botjes bevinden zich in hun peuterpuberteit, er gaat nog wel eens iets mis. Naar aanleiding van de gesprekken die ze voeren, probeer ik ze te verbeteren. Spreek ze gerust aan, wellicht vindt u uw woorden nog eens terug in de poëzie.
A Twitterbot remix of "This is a Picture of Wind: A Weather Poem for Phones” by J. R. Carpenter.
The sky clears its schedule. Gentle breezes. Lavender in need of attention. The river brings along a novel. #thisisapictureofwind
Late summer thunder. Heat rising out of nowhere. Elegiac, but we’ll take it. #thisisapictureofwind
Every hour, this bot draws language from wikiHow, repackages and recontextualizes it as a sexting message, and tweets it. Part of its process is to add pronouns “I,” “you,” or both to the instructions and actions described, in addition to prefacing each tweet with “sext.” Its output invites readers to interpret bland, utilitarian language metaphorically because it’s conceptually framed as sexting. The scenario of people sending sexually explicit messages back and forth, describing things they are doing to their bodies, contrasts sharply with the step-by-step instructions common to wikiHow, resulting in surprising and humorous results. Follow this bot on Twitter to learn many new euphemisms for sexual acts and the expressive potential of conceptual reframing. (Source: Editorial Statement from the works collection site)
Every three hours, this bot tweets a generated text field composed of blank spaces and unicode characters that can be interpreted as stars or other celestial bodies, particularly when conceptually framed by the account’s title. Its artistic output has become very popular, rapidly attracting over 70,000 followers and with each tweet being favorited and shared over 300 times. While this project would seem to be more of a visual art than literary bot, consider that it is not generating images, but sequences of characters, spaces, and carriage returns. It is using the materials of writing in the tradition of ascii art and its results are so evocative that it has even inspired a spinoff bot @tiny_astro_naut. Follow this bot to become to explore its tiny endless expanses. (Source: Editorial Statement from the works collection site)
witter bot, sample output available
Tiny Crossword is a daily game played publicly on Twitter. The bot posts a procedurally-generated three-word puzzle at noon PST. Players (any Twitter user) can @-reply with their proposed answer. After two hours, the bot posts the solution & credits the first player to have solved it. Twitter's constraints were designed for succinct handwritten messages, but bots explore what else can be expressed within those limits. The goal of this bot was to make a game that could fit into a tweet (117 characters with an image). Crediting the winner publicly also fits Twitter's form, where @-mentions can be a sign of admiration & prestige. Most bots generate content by taking a random walk through a large corpus. For Tiny Crossword, the corpus is Simple English Wikipedia; its brevity & plain language afford short puzzle clues. New puzzles are generated using up-to-date terms & concepts with no additional designer input.
The Way Bot scrapes Twitter for tweets that contain the phrase “I like it when.” After removing all the hashtags, special characters, and identifying material, the bot confirms that it has not encountered the expression before and then stores the statement in a database in order to tweet a fresh comment every few minutes. By removing specific markers of identity and filtering hateful or offensive language, what comes through is a more basic (and generally positive) expression of human feeling. The generative engine for The Way Bot was originally created to harvest a large amount of sentences starting with “the way that” and assemble them into the 99-page novel, The Way That I’m Crying so Hard I Have to Gasp For Air, by Eli Brody and 5,134 friends, a work submitted to NaNoGenMo (National Novel Generation Month) in 2013. Whereas reading the litany of statements out loud emphasizes The Way Bot's anaphoric poetics, embedding this account within a user's Twitter stream produces a lyrical refrain, a musical chorus that endlessly transmits a beacon of humanity’s collective unconscious. (Source: Editorial Statement from the works collection site)
Twitter bot, built in Ruby. Sample output available
This bot draws snippets of positive reviews from Rotten Tomatoes (a film and TV review aggregator), changes the director or producer’s name to a Fox News anchor or personality, and tweets it every two minutes. This Twitter account and bot were produced by The Colbert Report as a response to the news that Fox News publicists had thousands of fake social media accounts to try to spin any postings or comments against their news channel. By recontextualizing praise for film and television performances, narrative, and directorial style, as well as adding the #PraiseFOX hashtag, this overwhelmingly frequent, positive praise comes across as ironic and absurd. Its output also serves as a kind of subtweet because whenever anyone searches for one of the Fox News personalities on Twitter they’re likely to get many “Real Human Praise.” Following this bot may prove to be too much for readers because its frequent endless tweeting will certainly accelerate the current in your Twitter stream.
This bot draws from a NOAA (National Oceanic and Atmospheric Administration) buoy which gathers oceanic and atmospheric data and mashes it up with text from Moby Dick. The buoy became unmoored on March 10, 2013 and was set adrift—still transmitting data—in the Pacific Ocean until found on November 4, 2015. The result of combining snippets of live data from this floating bot with text from Moby Dick grounds its maritime language in a real and changing yet geographically distant and indeterminate present. Surf this bot’s poetic wanderings to explore real and imagined seas. (Source: Editorial Statement from the works collection site)
Twitter bot, sample output available
This bot is a tool designed to help readers explore the Electronic Literature Collection, Volume 3. Created with Cheap Bots, Done Quick!, a free bot hosting service powered by Tracery, an intuitive JavaScript library developed by Kate Compton, PhD. The bot currently tweets a suggested work from the ELC3 every 3 hours, linking to the work and adding the #ELC3 hashtag. Its Twitter account also compiles two lists of bots: a complete list of its 11 bots and one without Real Human Praise, which posts too frequently to allow readers to appreciate the other bots. Future development of this bot will include random suggestions based on ELC3 metadata, such as keywords, language, location, year, and we may even add some interactivity so it can respond to queries. In the meantime, follow this bot to receive suggestions of works to explore in the ELC3.
