What percentage of OpenAlex articles rely on web scraped affiliations data?

What percentage of journal articles in 2025 contain harvested affilaition data is affiliations are restored in OpenAlex?MethodologyWe know that as at the end of 2025 have dropped precipitously, both due to increased website protections and a bug that stopped the new harvesting code in OpenAlex from taking effect.To get a rough idea of the percentage of records that get their information from harvesting, we can take the difference in precentatge of publication records from 2023 - a pre AI protection year, and compare it with the percentage of records by publisher that have affiliations in 2025As OpenAlex is actively refreshing its affiliations data by harvesting, we use the last snapshot taken in 2025These numbers were calculated on Thursday 5th Feb 2026