Musk, who originally said part of his plan to own Twitter was to “beat the spam bots,” has recently accused Twitter of “ lie about the number of bots on his platform, and has argued that he should be able to walk away from the deal if Twitter doesn’t provide the information that necessarily backs up its publicly reported estimates. Twitter (TWTR) has sued Musk in an attempt to force him to close the deal.
from Musk response to Twitter’s lawsuit made public on Friday, says the billionaire’s team Twitter’s “fire hose” of tweets and Botometer to analyze the number of bots on the platform. Musk’s response claimed that based on his analysis, “fake or spam accounts” made up 33% of visible accounts on the platform in the first week of July, and about 10% of daily active monetized users. Twitter has long maintained in public filings that such accounts represent less than 5% of its daily active monetizable users.
Yang, one of the creators of Botometer, said he hadn’t heard from Musk’s team and was surprised to see the world’s richest man use his tools.
“To be honest, you know, Elon Musk is really rich, right? I assumed he would spend money hiring people to build an advanced tool or methods on his own,” Yang told CNN Business on Monday. Instead, Musk chose to use the Indiana University team’s free, publicly available tool.
Twitter has repeatedly argued that bots aren’t really relevant to the deal’s completion, after Musk signed a binding contract that doesn’t include bot-related carve-outs. Still, the company hit back in response to Musk’s response, noting that Botometer uses a different method than the company to classify accounts and “earlier this year, Musk referred to himself as very likely a bot.”
Botometer is indeed looking at the issue a little differently, Yang said. The tool does not show whether an account is fake or spam, nor does it attempt to make a different judgment about the intent of the account. Instead, it shows how likely an account is to be automated — or managed using software — based on various considerations, such as the time of day it tweeted, or whether it has declared itself to be a bot. . “There’s overlap, of course, but they’re not exactly the same,” he said.
The distinction highlights what could become a major challenge in the legal battle between Musk and Twitter: There is no single, clear definition of a “bot.” Some bots are harmless (and in somecases, even useful ones) automated accounts, such as those tweeting weather reports or news updates. In other cases, a human may be behind a fake or scam account, making it difficult to figure out with automated systems designed to take out bots.
Botometer gives a score from zero to five that indicates whether an account appears “human-like” or “bone-like.” Contrary to the characterization of Twitter, at least since June, the tool has rated Musk’s account as roughly one in five on the bot scale — indicating that there’s almost certainly a human behind the account. For example, it shows that Musk tweets quite consistently across all days of the week and that the average hours of his tweeting reflect a human schedule. (A bot, on the other hand, can tweet all night, during the hours most people sleep.)
But in many cases, Yang said, the difference between bone or not can be blurry. For example, a human can log in and tweet from what is normally an automated account. With that in mind, the tool is not necessarily useful for affirmative classification of accounts.
“It’s tempting to set an arbitrary threshold score and consider anything above that number a bot and anything below a human, but we don’t recommend this approach,” a commentary on the Botometer site said. “Binary classification of accounts with two classes is problematic because few accounts are fully automated.”
In addition, Twitter’s fire hose only shows accounts that are tweeting, so evaluating it would exclude bot accounts that, for example, simply want to increase the number of followers of other users — a form of inauthentic behavior that involves not tweeting, Yang said.
Musk’s legal team did not immediately respond to a request for comment on this story. But Musk’s response acknowledges that his analysis was “limited” due to the limited data provided by Twitter and the limited time he had to conduct the evaluation. It added that he continues to seek additional data from Twitter.
There’s private data from Twitter — such as IP addresses and how much time a user spends looking at the app on their devices — that can make it easier to gauge whether an account is a bot, Yang says. However, Twitter claims to have already provided more than enough information to Musk. It may be hesitant to hand over such data, which could pose a competitive risk or undermine user privacy, to a billionaire who has now said it no longer wants to buy the company and has even suggested starting a competing platform.