Up to 12% of Twitter users are bots, study says

Elon Musk's deal to buy the social media platform hinges on the accuracy of the number of bots

FILE- This April 26, 2017, file photo shows the Twitter icon on a mobile phone, in Philadelphia. Many companies large and small are updating their privacy policies and service terms to comply with upcoming European Union rules governing data and privacy. The new policy includes a few exemptions just for Europeans. Twitter says it may receive log data from websites that embed Twitter content. But the policy now states that Twitter won’t collect such data “from browsers that we believe” are in the EU and four countries linked to the EU by trade agreements, Iceland, Liechtenstein, Norway and Switzerland. (AP Photo/Matt Rourke, File)
Beta V.1.0 - Powered by automated translation

Twitter bots amount to up to 12 per cent of visits on the social media platform, a study has revealed.

The study by Israeli cyber-security company Cheq analysed 5.21 million website visits originating from Twitter, using more than 2,000 cyber-security tests to determine each user's authenticity.

It found that 11.71 per cent of all visits were driven by bots or fake users, including spam bots, scrapers, botnets, click farms and automation tools, as well as other forms of fake, fraudulent and non-human traffic.

The number of active bots on Twitter has come under the spotlight in recent weeks amid Elon Musk's deal to buy the company, and his attempt to buy it for less than the agreed $44 billion.

Mr Musk said the “deal cannot move forward” unless the microblogging company provides proof that less than 5 per cent of its users are fake.

In a regulatory filing earlier this month, Twitter said false or spam accounts represented fewer than 5 per cent of its “monetisable” daily active users during the first quarter.

“The data suggests that Twitter's bot problem is probably larger than 5 per cent,” said Guy Tytunovich, founder and chief executive of Cheq.

“Our study looked into users who came from Twitter to other websites. But if you consider that many bots don't click through to other sites and only stay on Twitter, then it seems very likely that bot traffic inside the platform itself could be significantly higher than 12 per cent.

“Ultimately, we are living in the era of the Fake Web, where bots, malicious users and automation tools make up a large portion of all web traffic, and this data supports what we are seeing out there.”

In April, Mr Musk had tweeted that one of his priorities would be to remove spam bots from the platform. He later tweeted that his acquisition of Twitter was temporarily on hold pending details on the amount of fake accounts on the social media platform.

He said he yet to see “any analysis” that suggests less than 5 per cent of Twitter accounts are fake.

“There is some chance it might be over 90 per cent of daily active users, which is the metric that matters to advertisers,” he said on Twitter.

Twitter chief executive Parag Agrawal took to Twitter to emphasise that fake accounts were less than 5 per cent of the total on the company's platform.

“We suspend over half a million spam accounts every day, usually before any of you even see them on Twitter,” he said in a tweet.

Twitter locks millions of accounts each week that it suspects could be spam, he said.

Meanwhile, research company Bot Sentinel estimated that 10 per cent to 15 per cent of accounts on Twitter are inauthentic, according to a Bloomberg report, while Cyabra, a research company with a different methodology, puts the percentage of inauthentic Twitter profiles at 13.7 per cent.

Updated: May 31, 2022, 12:28 PM