Bard or Bot?

Bard or Bot?

If, like me, you grew up in the ’80s and had sci-fi obsessed brothers who threatened you with Vogon poetry, you know that a poorly constructed verse can be painful. For some students, however, that agony extends beyond the purposely bad constructs of alien torturers to the mere mention of a ‘poetry unit’. Some people just don’t like poetry — whether it’s good, bad or somewhere in between.

If you’re looking for a fresh angle for teaching poetic structure and devices, or a new way to engage a particularly verse-averse class in the craft of poetry, consider introducing your students to poetry bots.

This week’s blog offers you a curated collection of sites that discuss non-human generated poetry. Some of the generators use algorithms to generate poems, while others use Google search terms or headlines from magazine editions.

Examining this kind of poetry requires students to consider the bot-poet’s techniques and structures to evaluate whether their ‘work’ should be considered genuine ‘poetry’. It is also a great opportunity to engage your talented coders, who might create a program to generate poetry of their own.

Reading through the automated texts, it also becomes very clear, very quickly that human poets reign supreme in the battle between the bards and the bots.

A few places you might begin

  • Bot or not includes a Turing-style test to assess whether readers can differentiate between poems written by humans and poems written by computers.
  • This Guardian technology article (by ) discusses the Google, Stanford University and University of Massachusetts research into the Recurrent Neural Network Language Model, an automated way of building sentences.
  • This article about the different types of computer-generated poetry online was posted in 2013 on The Verge, but has since been updated with many links to different types of generators.
  • Google poetics generates poems from the phrases that drop down as suggested search terms when you enter a few words.
  • This blog post by Swizec from 2012 also gives a great overview of some of the most significant experiments in automated text creation.

You can also create poetry using simpler forms of random generation. For example, write words and phrases on slips of paper, and give a handful to each student from a bucket.

Language-based fun and the study of serious poetic techniques aside, these generators are also an entry point into questions about the role of technology in communication, and whether the human brain is uniquely creative. If you feel so inclined, you could also broaden the discussion of computer-generated texts into a consideration of artificial intelligence. This could provide a lovely single-class diversion from the heavy thematic ‘what makes us human’ discussions that are necessary with the study of texts such as Gattaca, Never Let Me Go or Blade Runner.

Please note that all links above will take you to third-party sites. While we’ve done our best to vet the content for a general secondary-student audience, the links may include content or advertisements that are not suitable for all school contexts, or may have changed since we last viewed the content.

Best wishes and happy teaching, or as my Google-search poem would say:

Best friends
Wishes for friends
Happy Birthday
Teaching bubble.

Sandra Duncanson
Senior Editor

True Stories for Context Studies

August 29, 2016

Into the Wild

August 29, 2016