The philosopher Erik J Olsson leads a research group which, with the help of experiments, gets to the bottom of the phenomenon.

- It is a purely theoretically interesting field for me as a philosopher. I am interested in theory of knowledge. Then it is an urgent problem for society, says Erik J Olsson who is professor of theoretical philosophy at Lund University.

Erik J Olsson Erik J Olsson leads the research group consisting of philosophers, cognitive scientists and psychologists. They investigate a field that has grown rapidly but which, according to Erik J Olsson, is under-researched and where the methodology is under development.

The filter bubbles are considered by many to be a poison that erodes trust in society. They are considered to reinforce prejudice and are seen as a threat to democracy. Algorithms in search engines and in social media are often singled out as culprits in the drama.

The research group has focused on the search service Google, which in 2009 launched so-called personalized searches. Now the individual's search pattern would affect the results in the next search and the next and the next towards ever narrower results. If you click on hits with the heading climate change is a hoax, the next search for the word climate change would give you more hits about climate-skeptic information.

How did Google's personalizations go?

- We don't see much of an effect there. It doesn't seem like your searches are filtering future searches, as we've often been taught. It can be assumed that Google has absorbed the criticism here and played it down, says Erik J Olsson.

Could it instead be us who create the filter bubbles?

Yes, that seems more likely. In an experiment that Erik J Olsson and his collaborator Axel Ekström carried out in 2020, they followed 48 subjects, not only what they clicked on, but also their eye movements. The searches dealt with conflict-laden topics such as climate change and migration. Using eye-tracking, the researchers could see that the subjects' eyes were searching for results that confirmed their opinions.

- Even if they eventually clicked on other search results, their eyes first landed on results that confirmed their ideological beliefs.

This could, says Erik J Olsson, indicate that we carry a partially unconscious bias, which would be in line with psychological theories about so-called selective exposure. It may also be the case that the filter bubbles have an evolutionary survival value, that they strengthen the group community and that it can be dangerous to search for information that risks leading away from one's own flock.

- Most people probably want what they are looking for to be true. But in conflict between what is true or clashing with the group, it is more important to stick to the group.

An interesting secondary finding in the study was that the subjects who described themselves as liberal were more likely to pay attention to confirming search results than those who described themselves as conservative.

- The result is significant and is in line with other research. But our study is the first where you have actually been able to measure a connection between what you look at and ideological conviction, says Erik J Olsson.

The fact that it is primarily us who seem to create the filter bubbles, as the research so far shows, speaks against tougher regulations to deal with them, says Erik J Olsson.

But how harmful are the filter bubbles?

The researchers will examine that question in more detail in the next step of the project. However, it is clear that there is a tension between the filter bubble and the search for truth. And too much polarization creates mistrust towards those who have different opinions, says Erik J Olsson. The others are not only wrong, they are dishonest and lying.

- There is a risk that the filter bubbles lead to people not being able to talk to each other. It's not good. But there can also be a strength in having a strong group affiliation. For example, if you belong to a group that rebels against a dictator. Then a strong sense of belonging is a necessity, says Erik J Olsson.

TEXT: Thomas Heldmark