Daily Arxiv

This is a page that curates AI-related papers published worldwide.
All content here is summarized using Google Gemini and operated on a non-profit basis.
Copyright for each paper belongs to the authors and their institutions; please make sure to credit the source when sharing.

Empirical evidence of Large Language Model's influence on human spoken communication

Created by
  • Haebom

Author

Hiromu Yakura, Ezequiel Lopez-Lopez, Levin Brinkmann, Ignacio Serna, Prateek Gupta, Ivan Soraperra, Iyad Rahwan

Outline

This paper explores the impact of a generative AI-powered chatbot (e.g., ChatGPT) on human language and culture. Using hundreds of thousands of hours of YouTube academic lectures and podcasts, we use econometric causal inference techniques to analyze the spikes in usage of specific words following the launch of the chatbot. We find measurable increases in usage of ChatGPT’s preferred words (e.g., delve, comprehend, boast, swift, meticulous). We argue that this signals the beginning of a closed cultural feedback loop in which machines trained on human data exhibit unique cultural characteristics, which in turn transform human culture.

Takeaways, Limitations

Takeaways:
Generative AI chatbots show measurable impact on human language use.
It suggests the possibility of a closed cultural feedback loop through human-machine interactions.
It raises concerns about the loss of linguistic and cultural diversity and the risk of large-scale manipulation as artificial intelligence technology advances.
It highlights the need for further research on the evolution of human-machine culture.
Limitations:
It is difficult to accurately quantify the impact of ChatGPT.
The representativeness of the data used in the analysis needs to be reviewed.
Further research is needed to fully prove causality.
There is a lack of specific analysis of the risks of large-scale manipulation, including the loss of linguistic and cultural diversity.
👍