Kai-Cheng Yang, a Ph.D. candidate at Indiana University, “questioned the methodology used by Mr Musk’s team and told the BBC they had not approached him before using the tool,” a BBC article said today. Yang created the Botometer, a website tool that Elon Musk used to estimate Twitter’s spam percentage for a court filing. Yang reportedly said that Musk’s calculation “doesn’t mean anything.”
The first week of July’s Twitter firehose data, according to a Musk court filing on August 4, “shows that, throughout that span, fraudulent or spam accounts accounted for 33% of viewable accounts,” according to a Botometer analysis. Yang drew attention to the fact that the Botometer awards scores between 0 and 5, with 5 being the most “bot-like,” and that Musk’s court brief didn’t specify the point at which he decided to draw the line between a person and a robot.
Yang told the BBC, “You need to pick a threshold to cut the score in order to determine the prevalence [of bots].” “You will receive more bots and fewer humans if you lower the threshold from three to two. In light of the fact that Musk’s court filing “doesn’t make the details obvious,” Musk is free to act however he pleases. Therefore, the number has no significance for me “explained Yang.
In a previous interview with Yahoo, Yang stated that technically, “you may set whatever threshold you want and to obtain any result you want.” The Observatory on Social Media and the Network Science Institute at Indiana University collaborated on the Botometer project.
Musk was evaluated as a likely bot by Botometer.
As Twitter noted in a court filing, the Botometer itself once “suggested that Elon Musk’s own Twitter account was likely a bot, giving it 4/5.” According to reports, Musk’s Botometer score has varied between 0.5 and 4, indicating that on certain days, the tool views Musk as more human than a bot.
Twitter added that Musk’s team “has not disclosed what score they are applying to conclude an account comprises spam; consequently, their assertion is unverifiable.” Twitter added that a user account can be a bot even if it isn’t what the company regards as a false account or spam. Twitter cited bots that “report earthquakes as they happen or updates on the weather” as examples.
The Botometer can identify other varieties of genuine accounts as being likely bots. My verified Twitter account received a bot score of 3 out of 5 from The Botometer today, while the verified Ars Technica account had a bot score of 3.6 out of 5.
The FAQ section of the Botometer website advises against assuming all accounts beyond a specific threshold are bots. “Although it may be tempting to set a random threshold score and classify anything that is above it as a bot and everything that is below it as a person, this strategy is not advised. We think that examining the distribution of scores across a sample of accounts provides more useful information “In the FAQ.
Yang was astonished Musk hadn’t developed a superior tool.
Yang recently spoke to CNN and expressed astonishment that Musk chose to employ the Botometer rather than coming up with anything more exact. “You are aware that Elon Musk is extremely wealthy, right? I had believed he would spend money on employing individuals to create some advanced tools or techniques on his own “Yang said to CNN.
Use the Botometer as directed “The tool’s FAQ states that “people and robots have differing strengths when it comes to pattern identification. It should be used to supplement, not to replace, your own judgment. Some accounts that appear to a human observer to be bots or humans can trick a machine-learning system. Botometer, for instance, occasionally classifies “organizational accounts” as bot accounts. Similar to how an algorithm might correctly classify some accounts that are challenging for humans.”
After Musk attempted to back out of his agreement to pay $44 billion for the business, Twitter filed a lawsuit against him in the Delaware Court of Chancery. By contesting Twitter’s public declaration that less than 5% of its monetizable daily active users (mDAU) are spam or false, Musk has defended his attempt to violate the merger agreement.
Twitter argues that its estimations are accurate because they are “based on multiple human reviews (in duplicate) of thousands of randomly chosen accounts each quarter utilizing both public and private data.” Twitter claims that Musk does not have the authority to withdraw from the merger deal due to the volume of spam accounts.
According to Musk’s court statement, he has plans for a more extensive investigation of spam. “Defendants’ experts are continuing their analysis even now and, in anticipation of production of additional data by Twitter (including ‘private’ data that Twitter makes available to its human reviewers and contends is necessary to verify its reported less-than-5-percent spam and false user rate), intend to conduct a more comprehensive analysis and expect to present updated estimates and findings in expert reports and at trial,” Musk’s lawyers wrote.