A poem made by an AI read by me and a poem made by me read by AI’s.
The video – and the process behind it – explores the nature of co-creation of a person and an AI but also the nature of the uncanny valley. How does AI-generated art make us feel, and why, for most of us, the uncanny part are the visuals?
By just reading the texts without the video, what kind of feelings do the poems incite in you? Can “cold AI models” invoke feelings in us? Can or should we compare ourselves to AI? What is the role of humans in the creative process?
Tools used:
It all started by making a poem discussing two people based on a photograph. The poem was modified to examine the setting of poems made by humans (old) and AI (young).
With some usage of the “Speech Synthesis Markup Language”, Amazon polly was made to read the poem out loud
A Wav2Lip model was used to make the lip-synced video
A GPT-3 model was used to generate poems with the prompt: “A rhyming poem inspired by Emily Dickinson talking about two poets of differing ages. This is the poem:” By this prompt the model made multiple poems from which one was chosen.
This was read to a camera by Olli