Natural language processing and semantic change can help policymakers, public health authorities, journalists, and fact-checkers keep a finger on the pulse of the online health conversation.
In 2019, Meedan’s Digital Health Lab in collaboration with researchers at the Oxford Internet Institute, the Harvard T.H. Chan School of Public Health, and University of Michigan’s Department of Computer Science developed and tested approaches to help policymakers, public health authorities, journalists, and fact-checkers keep a finger on the pulse of the online health conversation.
By modifying semantic change detection methods, we found we could identify when health words were being used in new, unexpected ways. Many of these shifts were due to large external events, but a few were due to misinformation.
The research is detailed in an 8-minute video and will be discussed at Global Fact 7, the annual conference of the International Fact-Checking Network on Wednesday, June 24.
