The endless battle against spam accounts on Twitter has become a problem for new owner Elon Musk.
He later cited bots as a reason to stop buying social platforms. Now that the billionaire has completed the deal, he faces the task of fulfilling a promise to clean up the fake profiles that have plagued Twitter and haunted him long before he expressed his interest in the acquisition. ing.
Challenges come with high stakes. The number of bots is important because advertisers, Twitter’s main source of revenue, want a rough idea of how many people they’re actually reaching when buying ads. It is also important to stop bad actors from amassing a large number of accounts to amplify misinformation or harass political opponents.
“The big picture in my head is how do we make Twitter a better place for everyone,” said Emilio Emilio, a bot-counting expert who spent the summer investigating the mask issue. Ferrara said. The value of the platform as
To find out how bad the bots were, Musk hired Ferrara and other data scientists to investigate. At the time, he tried to prove Twitter was misleading the public when he said less than 5% of daily active users were fake or spam accounts. Musk could have argued that terminating the $44 billion contract was justifiable.
Ferrara, an associate professor of computer science and communications at the University of Southern California, said he wasn’t too concerned about whether Musk would end up owning the platform.
Instead, when Ferrara first told The Associated Press about his planned role as an expert trial witness for Mr. Musk, he hoped “any discoveries will help improve the platform.”
The question is what Mr. Musk will do with that information. Ferrara’s presentation — about 350 pages of analysis and supporting documents — is kept in top secret court documents, and he said he could not disclose his conclusions.
A former Twitter executive and his lawyer said Mr. Musk exaggerated the issue because of buyers’ remorse. An exact number is “nearly impossible” because the bot’s estimates are based on assumptions that can lead to bias, said researcher Filippo Menzer, who did not work for either side of the controversy. rice field.
Mentzer, director of the Social Media Observatory at Indiana University, said, “No one knows exactly how serious this problem is. It’s not as bad as Musk said, and not as Twitter claimed.” I don’t think it’s good,” he said.
Many experts have also questioned Musk’s ability to easily make improvements, with Musk citing the use of algorithms to track and remove fake accounts and to “authenticate” real people. suggesting that it relies on implementing new means of
Ferrara was preparing to travel to the East Coast to testify in Delaware earlier this month. He said he would proceed with the acquisition in dollars. Closed on Thursdays.
Most legal experts didn’t think Musk had many lawsuits. The court’s chief justice appears likely to favor Twitter based on certain terms of the April purchase agreement.
However, according to Ferrara and other researchers hired by Musk’s legal team, Musk was not without his opinions on bots.
The analytics firm CounterAction, which works with Ferrara, said in a July 18th court-filed report that Twitter’s spam rate for monetizable accounts (accounts of value to advertisers) is at least 10%, It said it concluded that it could reach 14.2%. It depends on how the rate is measured.
The company’s founder and CEO, Trevor Davis, said the analysis was based on a “fire hose” of internal data Twitter provided to Musk, but the company did not provide additional data requested by Musk’s team. Rejected.
“We hope that access to withheld data will reveal that true spam rates are even higher,” Davis said in a prepared statement.
As a celebrity user with over 110 million followers, Musk has long been obsessed with Twitter spambots promoting cryptocurrency schemes. Some scammers have even opened accounts mimicking Musk’s name and likeness in an attempt to trick people into thinking he’s endorsing something.
Not all bots are bad. Twitter encourages the use of automated accounts that post weather, earthquakes, or lines of humor or classic literature. Twitter also allows anonymity, which protects free speech and privacy. Especially in autocratic regions. But that approach can make it harder to eradicate malicious fake accounts.
Ferrara first came to Twitter’s attention after it emerged in 2016 that Russia used social media to interfere in the US presidential election. It was when he led a research group that he estimated that 15% of his 9% of active English-language accounts on Twitter were bots.
In a blog post soon after, Twitter complained that such external research was “often inaccurate and methodologically flawed.” The company has repeatedly reported figures below 5% in its quarterly reports to the Securities and Exchange Commission, but warns that it could be higher.
Before the Musk acquisition, Twitter said it removed 1 million spam accounts every day. To calculate the number of accounts that are malicious spam, Twitter uses both public and private data, such as IP addresses, phone numbers, location information, and what happens when an account is active. Review thousands of randomly sampled accounts.
But over the past few months, Musk and Twitter have been battling over that methodology. Twitter uses a metric called mDAU for monetizable daily active usage.
“This is literally the metric they invented. You can’t contrast or compare that metric to any other service,” Ferrara said.
When Musk first publicly asked questions about the number of bots after agreeing to buy the company, Cyabra, another Israel-based company, said it had an answer.
“The elusive number you’re looking for…we have it. It’s 13.7%,” the company tweeted on May 17, referring to Musk’s Twitter handle to get his attention. flagged.
Cyabra’s machine learning technology scans dozens of social media profiles to track behavioral patterns and try to find those behaving like humans. Guesswork like that can go awry, but the tweet caught the attention of people close to Musk, if not billionaires.
Cyabra CEO Dan Brahmy said the company began working with Musk’s campaign by the end of May. Whatever the true count, he said it would not be an easy problem to solve.
“Some bots are definitely villainous,” says Brahmy. “The trade-off is between having very high sign-up standards and information security, and being very open-minded in a way, promoting free speech and creativity.