According to a new study that looked into how fake and conspiracy news flourished on the micro-blogging site Twitter both before and after the 2016 US presidential elections, Twitter still remains a breeding ground for accounts spreading fake news.

A study conducted by the Knight Foundation said on Thursday more than 80 per cent of the Twitter accounts linked to spread of disinformation during the 2016 election are still active and these accounts continue to publish more than a million tweets on an average day.

The researchers used tools and mapping methods from Graphika, a social media intelligence firm to study more than 10 million tweets from 700,000 Twitter accounts that linked to more than 600 fake and conspiracy news outlets.

Twitter, along with other social media platforms including Facebook came under the intense scrutiny of US policymakers for their failure to stop the spread of misinformation on their platforms during the 2016 election.

Twitter since then has stepped up its efforts to curb the spread of divisive messages and fake news on its platform.

Twitter had earlier this week announced that it will delete fake accounts engaged in a variety of emergent, malicious behaviours to further protect the integrity of elections.

As platform manipulation tactics continue to evolve, the micro-blogging platform said it is expanding rules to better reflect how it identifies fake accounts and what types of inauthentic activity violate its guidelines before the US mid-term elections in November.

As part of the new rules, accounts that deliberately mimic or are intended to replace accounts that were previously suspended for violating rules may be identified as fake accounts, Twitter said.

The Knight Foundation study found more than 6.6 million tweets linking to fake and conspiracy news publishers in the month before the 2016 election.

“Machine Learning models estimate that 33 percent of the 100 most-followed accounts in our postelection map — and 63 percent of a random sample of all accounts — are “bots,” or automated accounts,” the study said.

“Because roughly 15 per cent of accounts in the postelection map have since been suspended, the true proportion of automated accounts may have exceeded 70 per cent,” it added.

(With agency inputs)