A survey by Litres showed that 54% of digital authors use generative models in their work on a piece. Authors most often use neural networks to check grammar, spelling, and punctuation (20%), search for information (19%), and create illustrations for the piece (15%).
Among those who do not use AI, 53% of respondents have a negative attitude towards its use, 42% are neutral, and 5% are positive. The main reason for refusal was the lack of advantages of using AI in literature (30% of respondents think so), although 23% note the usefulness of analyzing big data, and 15% believe that neural networks save time. AI users note that neural networks help improve the quality of the text (24%).
Among the main risks, respondents cite the loss of original style and author's voice, the development of dependence on generative models, the loss of literary skills, as well as factual inaccuracy.
As for labeling, 72% of authors using AI believe that it is not necessary to indicate the fact of using neural networks in a book, while 81% of those who do not use AI believe that the reader should be notified about the use of generative models.
Read more on the topic:
Yandex to Present Joint Works of Neural Network and Human
Artificial Intelligence Revived Kustodiev's Paintings at the Tretyakov Gallery
Opera Singer from Russia First in the World to Sing with Artificial Intelligence