Twitter study reveals its algorithms amplify political tweets from right-leaning users

• Twitter conducted a study on whether its algorithmic timeline amplified any political ideology or group.
• It found that tweets from the political right are amplified more than the political left.
• This occurs more in its algorithmic timeline as compared to the reverse chronological timeline.

Twitter has revealed findings of its research conducted on the algorithm it uses for the home feed. It analysed political content on the platform and whether a certain political ideology was amplified by its algorithms. The study found that Twitter indeed amplified tweets posted by right-leaning accounts as compared to those from the political left.

The research involved analysing millions of tweets between April 1 and August 15, 2020, and from accounts operated by elected officials in Canada, France, Germany, Japan, Spain, the United Kingdom, and the United States. It also included tweets containing links to news articles shared by people on Twitter.

The comparison was made between Twitter’s algorithmic timeline and the reverse chronological timeline. Twitter lets you choose between ‘top tweets’ and ‘latest tweets’. Top tweets show tweets from accounts you follow and also recommended tweets that may be of interest to you. Latest tweets show the most recent tweets as it happens on the timeline.

The research’s findings revealed that tweets about political content from elected officials do get amplified in the algorithmic timeline as compared to the reverse chronological timeline. A big revelation here was that in the six countries excluding Germany, Twitter’s algorithm showed more tweets from the political right than the political left. This was the case for right-leaning news outlets too which were pushed more than left-leaning news outlets.

Twitter clarified that algorithmic amplification would happen but if there’s preferential treatment then “further root cause analysis is required in order to determine what, if any, changes are required to reduce adverse impacts by our Home timeline algorithm.”

The research was conducted by Twitter’s ML Ethics, Transparency and Accountability (META), and it’s part of Twitter’s initiative to improve its machine learning (ML) algorithms. Twitter also conducted studies on whether its image cropping algorithm is gender and racial biased, and it found that images of white individuals were favoured over black individuals. Following the results of its research, Twitter said it updated its system to show more accurate image previews.

SEE ALSO:

Snapchat’s upcoming family centre feature will allow parents to see who is chatting with their teenage children
Apple could drop ‘Air’ for next MacBook with M2 chip, iMac-like colours, and MagSafe