Homogenization Effects of Large Language Models on Human Creative Ideation
2024-08-26 18:02:38.113096+02 by Dan Lyke 0 comments
Homogenization Effects of Large Language Models on Human Creative Ideation
<blockqutoe>Coupled with evidence that ChatGPT users exhibit greater fluency, flexibility, and elaboration than users of an alternative CST, these results suggest that current general-purpose instruction- tuned LLMs (such as ChatGPT) are capable of functioning as useful CSTs by enabling the rapid enumeration of relatively obvious possibilities that users might otherwise fail, or take longer, to consider. However, these systems are not currently well-suited to helping users develop truly original ideas.
https://doi.org/10.1145/3635636.3656204
Via Mark Gritter, who links to some additional resources used in the paper.