The limits of verification in preventing the spread of false information on networks

Abstract

The spread of false information on social networks has garnered ample scientific and popular attention. To counteract this spread, verifi cation of the truthfulness of information has been proposed as a key intervention. Using a behavioral experiment with over 2000 participants we analyze individuals' willingness to spread false information in a network. All individuals in the network have aligned incentives, making lying attractive, countering an explicit norm of truth-telling that we imposed. We investigate how verifying the truth, endogenously or exogenously, impacts the choices to lie or to adhere to the norm of truth-telling, compared to a setting without the possibility of verifi cation. The three key take-aways are: (i) veri fication is only moderately effective in reducing the spread of lies; its effectivity is (ii) contingent on the agency of individuals to seek truth, and (iii) the exposure of liars, and not only the lies told. These suggest that verifi cation is not a blanket solution. In order to enhance its effectivity, it should be combined with fostering a culture of truth-seeking and with information on who spreads lies, not only on the lies told.