The biased reality of online media - Using stereotypes to make media manipulation visible

Summary

The application of artificial intelligence to digital media enables easy manipulation of the information that reaches us (fake news, echo-chambers). This opens up opportunities to control our beliefs and, thereby, our actions (Cambridge Analytica), which threatens personal freedom and democracy.

To deal with these risks we need to know more about how what we read on (social) media affects our beliefs and actions. In this interdisciplinary project we will use knowledge from linguistics and psychology to build a cognitive model that predicts how media sources affect stereotypical beliefs: general beliefs about characteristic features associated with (social) groups. Stereotypes play a central role in decision making, as they allow us to make predictions. This makes stereotypes a key factor in how media can affect our thoughts and behaviour. What you read about Brexit will inform your stereotypes of Great Britain/the EU and, thereby, influence your voting behaviour.

Our model will make concrete predictions for which stereotypes are endorsed by consumers of particular media. We will test the model using the news coverage of the US presidential elections and compare the predictions with stereotypes voters associate with the competing parties and candidates.

The knowledge and technology this project will deliver will help government and industry to formulate and enforce benchmarks to control information filtering. Additionally, we will utilise the developed technology in the realisation of an online tool, the “Bias Barometer”, that predicts, given a media corpus, which stereotypes it induces.

Details

Project number

406.DI.19.059

Main applicant

Dr. K. Schulz

Affiliated with

Universiteit van Amsterdam, Faculteit der Geesteswetenschappen, Departement Filosofie

Team members

SAM Hogenboom MSc, OD van der Wal MSc

Duration

01/02/2020 to 31/01/2024