Dice Question Streamline Icon: https://streamlinehq.com

Authenticity of observed U.S. declines versus database artifact

Ascertain whether the observed declines in U.S. AI policy publication counts and preprint citation rates around COVID-19 and post-ChatGPT are genuine trends or artifacts arising from reliance on Web of Science data.

Information Square Streamline Icon: https://streamlinehq.com

Background

The U.S. shows dips during COVID-19 and after ChatGPT’s release, unlike Europe and South Korea. The authors note that coverage differences across bibliometric platforms may explain this divergence.

Validating the trend requires integrating alternative data sources to assess whether the dips are real or a function of Web of Science’s indexing scope.

References

Robust data for the U.S. response to COVID-19 results suggest there may be other web resources (outside Web of Science) where a greater majority of European and UK, as well as Korean scholars, are more well represented, and therefore, it is difficult to determine if this is an actual trend or an effect of our use of only Web of Science data.

— Shifting 'AI Policy' Preprints and Citation Trends in the U.S., U.K and E.U., and South Korea (2015-2024) (2510.16477 - Suh et al., 18 Oct 2025) in Conclusion