Isabelle Levent

There are a couple reasons why Wordcraft may have struggled with style and voice... Another reason could have been limitations of the underlying model. LaMDA and other similar language models are trained to be most confident on the kind of text they see most often–typically internet data. However, professional creative writers are usually writing

The “story seeds” control allowed writers to experiment with using Wordcraft to generate ideas at the very beginning of the writing process. MT described the “story seeds” control as giving her “a place on my computer to go that looked like a blank page but did not behave as such.” E

Now none of this is meant to say that I think programmers, artists and engineers have no responsibilities when it comes to the outputs of machine learning models. In fact, I think we bear responsibility for everything these models do. (I never, for example, attribute authorship to a program or a model. If I publish the results of a text generator,

For a computer to make a subtle combinational joke, never mind to assess its tastefulness, would require, first, a data-base with a richness comparable to ours, and, second, methods of link-making (and link-evaluating) comparable in subtlety with ours. 

However, we often found that it was the unexpected differences between the prompt and the generated image’s interpretation of it that yielded new insight for and excitement from participants.