Artistic Works

Promises

NLP can be used to generate seemingly creative or poetic text that has no human input but relies on prior creative work. A simple prompt “Write a poem” into GPT-3 or a comparable transformer would produce a seemingly creative output with no effort from the user.
It can generate content that is distinct enough to avoid a copyright violation, but sufficiently similar to the original to serve as an analogue.

Already existing examples of creativity NLP systems include “VersebyVerse” that produces poetry inspired by classic American poets.
NLP AI applications can be used to stimulate creativity and help generate ideas for creative writing by suggesting next sentences, story plots or phrasing.

If such applications were used at scale, artists and creatives might have even less opportunities to make a living, which would seriously impact the viability of the sector.
Legal concerns of NLP systems directly reproducing copyrighted material from the training data is subject to legal discussions.

Because NLP relies on a statistical computational process without any understanding of meaning or semantics, void of intention, and disconnected from action and responsibility, texts generated cannot be considered on a par with language produced by human speakers.
Some cultures are massively overrepresented in the training data, which means the LLPs performance for minority languages, dialects and cultures is much worse. What implications does the limitation have on the users that are left out?
If a NLP system assumes the creative expression of a person, living or deceased, how can we assure that their dignity is respected? Do the authors have control over how their work is used?