The methodology here appears to have been doing a keyword search among some 88k papers published between 2012 and 2020. It's obvious to me that humans have cau

The methodology here appears to have been doing a keyword search among some 88k papers published between 2012 and 2020. It's obvious to me that humans have caused changes, but this methodology seems awfully lazy.