Twitter Says Musk’s Spam Analysis Used a Tool That Called His Own Account a Bot

Twitter yesterday denounced Elon Musk’s response to the company’s lawsuit in a 127-page filing with the Delaware Court of Chancery stating that Musk’s claims “contrary to the evidence and common sense.” Twitter’s lawsuit also said Musk’s spam analysis was based on a tool that once called his own Twitter account a likely bot.

“According to Musk, he – the billionaire founder of multiple companies advised by Wall Street bankers and lawyers – was tricked by Twitter into signing a $44 billion merger agreement,” Twitter wrote. “This story is as implausible and inconsistent as it sounds. And it’s just that — a story, conceived in an effort to escape a merger deal that Musk no longer found appealing once the stock market hit the market. — and with it his enormous personal wealth — depreciated in value.”

Twitter’s submission was in response to Musk’s defense and counterclaims, which were filed last week but were not immediately made public because Twitter was given time to request redactions. Twitter apparently chose not to do any editing.

Musk’s counterclaims, which are based on distortion, misrepresentation and outright deception, do not change anything. Musk has signed and is obligated to honor the merger agreement, Twitter wrote.

Twitter: Musk used tool that called his account a bot

Musk has defended his bid to break the merger agreement by questioning Twitter’s public disclosure that less than 5 percent of its revenue-generating daily active users (mDAU) are spam or fake. The Twitter court said Musk’s own analysis used a publicly available website to find “that fake or spam accounts make up at least 10 percent of Twitter’s daily active users,” but “Musk doesn’t measure the same as Twitter and doesn’t even use Twitter.” same data as Twitter.”

Musk can only make a higher estimate by running a dataset not limited to or including mDAU through a generic web tool that pointed his own Twitter account to a likely ‘bot’. The result is a distortion that Musk hopes will nevertheless will cause a lot of commotion,” said Twitter.

In particular, Musk used “an internet application called the ‘Botometer’, which uses different standards than Twitter and which Musk himself identified earlier this year as highly likely to be a bot,” according to Twitter.

The Botometer website is a project of the Observatory on Social Media and the Network Science Institute at Indiana University. Citing an article from the May 2022 Protocol, the Twitter court said that “The Botometer indicated that Elon Musk’s own Twitter account was likely a bot and scored it 4/5.”

Musk’s Botometer Score Now More “Human”

This morning, Botometer Musk’s account gave a rating of 1.2 out of 5, indicating that as of today, Musk is more “human” than bony. Protocol’s article in May noted that Musk’s account got wildly different Botometer scores overnight, saying the tool “highlights how difficult it is to identify bots, especially using only public data.”

As Twitter’s court noted, “the Botometer’s own FAQ website warns that ‘bot detection is a difficult task’ and that if it was ‘easy to do with software, there would be no bots — Twitter would have caught them and banned them already’.” !'”

Twitter’s request described other issues with Musk’s reliance on the Botometer as follows:

So the Botometer doesn’t even pretend to apply Twitter’s definition of a fake or spam account. Some bots (such as bots that report earthquakes as they happen or weather updates) are often helpful and allowed under Twitter’s platform manipulation and spam policy, which Twitter respectfully refers the court to. In addition, Defendants have not indicated what score they use to conclude that an account is spam; their claim is thus not verifiable.

Musk already lost a key ruling when Chancellery Judge Kathaleen McCormick sped up the process and denied Musk’s request to postpone it until February 2023. The five-day trial now begins on October 17.