Live presentation

Asynchronous discussion


Disinformation and political polarization: A study of the impact of political beliefs in the ability to detect “fake news”

Laouen Mayal Louan Belloli

  • CABA,
  • Argentina
  • Laouen Belloli ¹𝄒²
  • , Joaquín Navajas ²𝄒³
  • , Nira Dinerstein ⁴
  • , Ariel Merpert ⁴
  • , Laura Zommer ⁴
  • , Andrea Goldin ²𝄒³
  • 1 Laboratorio de Inteligencia Artificial Aplicada, Depto. de computación, FCEyN, UBA–CONICET
  • 2 Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET), Ministry of Science, Buenos Aires, Argentina
  • 3 Laboratorio de Neurociencia, Universidad Torcuato Di Tella
  • 4 Chequeado

The use of social media is growing, and the spread of ‘fake news’ with direct impact on political decisions has raised to new levels. Social media (and their non transparent algorithms) are the main organisations to be pointed out as responsibles for this issue, and they started to implement different strategies on their platforms to reduce disinformation. One of the proposed mechanisms is not allowing people to share a news before a certain delay, in order to reinforce inhibitory control, that, in theory, permits a more rational analysis of the news.

Polarization is known to affect people’s judgement. Since its origin, Argentina has had a strong political crack, nowadays represented by two of its major political parties: Kirchnerismo and Macrismo.

In this work, we studied the impact of Argentinian political polarization on the subjects’ ability to detect disinformation and misinformation, and the use of delay times to improve their performances. For this purpose, we conducted an experiment where 715 subjects classified a set of news as ‘fake’ or ‘real’ with or without a previous delay time. We found (1) a strong confirmation bias when the news disagreed with subjects’ political beliefs, and (2) that adding a delay not only does not help to deal with that bias, but worsens the detection of real news when the news is aligned with the subject’s political beliefs.