Social media content does not seem relevant
People who spend hours looking at photos and videos on Instagram don't necessarily think social media content is important. The voting booklet clearly wins out here, as a study shows.
Whether it's Google, YouTube or Instagram − it's often not us who decide what we see on the Internet, but algorithms. This is why many fear that our thoughts will be increasingly manipulated as a result. However, a study funded by the SNSF now shows that people do not necessarily consider these channels to be highly relevant, even though they may spend a lot of time on them.
A research team from the University of Zurich determined the degree of relevance that over 1,000 respondents assigned to information from various sources – not only online content, but also offline alternatives such as newspapers, television or family. "These are often disregarded in direct comparison," says media researcher and co-author Tanja Rüedy.
But as it turns out, such traditional sources are still important, for example in shaping political opinion. "We were surprised that nearly 70 per cent of respondents considered the Swiss voting booklet to be relevant," says Rüedy. Television, radio, newspapers and personal contacts performed similarly well. By comparison, only about 20 per cent considered the content of algorithm-driven social media relevant. Comparable results were obtained in other areas such as health and purchasing decisions.
Search engine results more relevant
However, the researchers cannot rule out the possibility that the respondents deliberately devalued the relevance of social media because there is a lot of negative publicity about it – for example in connection with fake news and filter bubbles.
They also didn't expect that search engine recommendations would be assigned considerably more relevance than social media. Perhaps the fact that search engines also use algorithms is less well known – though this did not form part of the study. "In general, Internet users are often unaware of where algorithmic selection is used. This awareness should be promoted more," says Rüedy.
The findings thus offer up another piece of the puzzle in understanding the impact of algorithms on our lives. "The ultimate goal is to correctly assess the associated risks to our society," says lead author Michael Reiss.