Thursday, October 22, 2020 - 4:00pm to 4:30pm
We in the mainstream media, stung by unfair tarring with the "fake news" label, are hypersensitive to the growth of actual misinformation online. And it's galling to see our Twitter feeds and the comment streams beneath our stories polluted by bigoted, chauvinistic, and sometimes threatening snark and vitriol. But new research suggests this front-row view to the downsides of free speech may have distorted our perspectives on these problems. Powerful observation biases skew what we as science writers see, compared to the general public.
In this session, researchers who have assembled big data sets to study these questions will present more objective views of the consumption of misinformation by the public and of how people respond to organized hate speech. The results of their quantitative analyses of large-scale, longitudinal studies challenge conventional wisdom about these phenomena.
Two scientists from the Santa Fe Institute who have used AI to classify nearly 200,000 Twitter conversations, spanning four years of activity on political accounts and large news sites in Germany, will share answers to a crucial question: is it better to ignore trolls or to combat them with counter-speech?
Then a data scientist at Penn who previously worked at Yahoo and Microsoft Research will update us on what his team's study of media consumption among large, nationally representative audiences in the U.S. reveals about how often Americans are exposed to truly fake news—and news of all kinds. The problems of ensuring an informed electorate, this research suggests, are quite different from what most of us have assumed.
- Mirta Galesic (Speaker) Professor, Santa Fe Institute
- Joshua Garland (Speaker) Applied Complexity Fellow, Santa Fe Institute
- Duncan Watts (Speaker) Stevens University Professor, Annenberg School for Communication, University of Pennsylvania
- W. Wayt Gibbs (Moderator) New Horizons in Science program director, CASW