More than 80% of accounts known to have tweeted links linking to fake and conspiracy news publishers are still active according to a study on how fake news spread on Twitter by the Knight Foundation.
The Knight Foundation used tools from social media intelligence company Graphika to track and analyze 700,000 Twitter accounts who tweeted content published by 600 conspiracy and fake news outlets more than 10 million times.
According to the analysis, automated accounts designed to spread disinformation (also known as Twitter bots) are behind the vast majority of all fake news spreading on Twitter.
Moreover, most of them have been clustering in groups designed to run coordinated campaigns push the same fake news stories for covering the possible number of targets.
Knight Foundation's research team discovered that out of all accounts that peddled disinformation material before and after the 2016 election, more than 80% are still active and pushing out more than a million tweets per day.
Nine out of the ten most linked fake news websites during 2016 elections are still active today
The study raises questions about Twitter's capabilities and efforts in detecting and blocking accounts known to have been active in fake news disseminating campaigns although there is enough proof of their illegal activity.
Furthermore, the researchers found out that ten fake and conspiracy news "providers" were being linked by 65% of all automated Twitter accounts spreading disinformation on the platform, while the top 50 were the receptors of 89% of all fake news links.
The big issue is that nine out of the ten conspiracy and fake news outlets that were in the top 10 most linked websites by disinformation bots before and after the 2016 election are still being actively linked and used to spread false information on various subjects.
CEO Jack Dorsey said in a testimony in front of the United States Senate Select Committee on Intelligence that Twitter had to disable between 8.5 million and 10 million accounts every week in 2018 due to spam spreading and automation suspicions.
Still, according to the Knight Foundation's analysis, not enough.