A group of in-house researchers found that content on the right tends to be amplified more than content on the left. Debate continues (including institutional) about the impact of algorithms on society and how to regulate them
The algorithm is not a problem in itself. Computational amplification – in this case, the process by which an algorithm learns to publish a certain type of content more than the standard – not even. But when a social platform becomes an essential component of democratic life, when messages from politicians and newspaper articles pass through it and the algorithm chooses which one to serve users, the impact that unbalanced amplification can have becomes immediately apparent.
This is what Twitter researchers were thinking when writing their latest news Link on arithmetic amplification of politics on their platform. They studied over three months in 2020 in seven countries (Canada, France, Germany, Japan, the United Kingdom, Spain and the United States). They focused on publications from local politicians and major news outlets, while excluding non-political content such as recipes and sports, relying on two independent websites, All sides NS media advertising sourcesto determine their political leanings.
The authors were not intent on defining the vision of a particular politician or newspaper, but rather they were intent on understanding the effects of algorithmic amplification on public discourse. Not only did they find that content posted by incumbent politicians is on average more amplified, but they also found that in all the countries considered, with the exception of Germany, posts from the right receive more amplification than those from the left. The same goes for titles that tilt to the right.
“In this study, we determine what is happening: Certain political content on the platform is amplified,” the researchers wrote. “Determining why these observed patterns occur is more difficult, because they are a product of interactions between people and the platform.” Twitter wants to identify and mitigate any imbalances that emerge, but “we need more root cause analysis to determine if changes are needed to minimize negative impacts” of the algorithmic selection system and what the changes are.
Here the rigorous empirical work of Twitter researchers necessarily limits the most smoked field of sociology. For many commentators, the reason lies in the right’s communication style, which is generally more emotionally charged and more likely to elicit an “intuitive” response. It has been widely proven that high emotional impact content creates more engagement and encourages the user to interact with it; L ‘link It is actually one of the factors that increases the value of content in the eyes of social algorithms.
The link between the right content and more link It remains (necessarily) public discourse. But the fact remains that Twitter’s findings are in line with other studies and research. The same Facebook data, analyzed through Mark Zuckerberg’s social media tools, they revealed The best performing political content in the United States comes primarily from personalities and newspapers on the right of the political spectrum. Observable trend 10 pages receive more link every day in the United States.
The effects of algorithmic amplification are at the heart of the findings whistleblower ex Facebook, Frances HoganWhich is touring institutions on both sides of the Atlantic to contribute to the work of politicians at a critical time. Both the US and the EU are trying to regulate the activity of social platforms and some proposals also assume that proprietary algorithms are “opened”, in the name of transparency. But even Facebook itself has urged regulators to set clearer rules, stressing that a series of choices with strong political value should be in the hands of institutions, not corporations.
“Introvert. Avid gamer. Wannabe beer advocate. Subtly charming zombie junkie. Social media trailblazer. Web scholar.”