Large language models reduce public knowledge sharing on online Q&A platforms (PNAS 2024)

This study examines the impact of ChatGPT, a large language model, on online communities that contribute to public knowledge shared on the Internet. We found that ChatGPT has led to a 25% drop in activity on Stack Overflow, a key reference website where programmers share knowledge and solve problems. This substitution threatens the future of the open web, as interactions with AI models are not added to the shared pool of online knowledge. Moreover, this phenomenon could weaken the quality of training data for future models, as machine-generated content likely cannot fully replace human creativity and insight. This shift could have significant consequences for both the public Internet and the future of AI.

Large language models reduce public knowledge sharing on online Q&A platforms

Abstract. Large language models (LLMs) are a potential substitute for human-generated data and knowledge resources. This substitution, however, can present

OUP AcademicMatjaz Perc

Note: at Ask for the moon, we work to alleviate this effect https://home.askforthemoon.com/

Bonus link: LLM watermarking https://pierrefdz.github.io/publications/threebricks/