With his expressed concern about fake accounts on Twitter†
Elon Musk appears to be grabbing legal straws in an effort to cancel his promise to buy the social networking company for $54.20 a share, or at least pay less for it. But his gamble has shed light on a real plague of online businesses and their users.
Counting the autonomous accounts that mimic real people is just as slippery as valuing companies. A 2020 study by Adrian Rauchfleisch and Jonas Kaiser, looking at thousands of Twitter accounts, including hundreds of verified politicians and “obvious” bots, found Botometer, the industry-standard learning algorithm trained to calculate the probability that an account is a bot. is, yields inaccurate scores leading to both false negatives and false positives.
Perhaps more disturbingly, the findings of a 2018 study published in Science that looked at about 126,000 stories — true and false — circulated on Twitter from 2006 to 2017 indicate that humans are more likely than robots. to spread false news.
Darius Kazemi, a computer programmer who spent ten years creating and studying bots, describes the difficulties in identifying bots as twofold: first, bots are measured by internal metrics that are usually not available to the public; second, the machine learning algorithms built to identify bots are informed by human assessment interviews, which are often wrong.
At the very least, that could be better than what Mr. Musk suggested: He said last week that his team would try to determine a more accurate percentage of Twitter’s bots by taking “a random sample of 100 @twitter followers.” He invited his followers to join the exercise as well. Mr. Musk recently estimated that fake users make up at least 20% of all Twitter accounts. In filing securities, Twitter has long estimated that fake or spam accounts represent less than 5% of the total number of active users, but has also said the actual number “could be higher than we estimated.”
When Twitter CEO Parag Agrawal tried to demonstrate the challenges with both internal and external bot identification on Twitter, Mr. Musk broke off the discussion with a poop emoji. It’s not even clear that a botless platform wouldn’t suffer somehow. Mr. Musk has likened Twitter’s bots to termites in a house, but they can also do some good: For example, a newspaper bot that automatically posts top headlines every day is an obvious public service.
People misidentify bots for a myriad of reasons, according to Mr. Kazemi. Active lurkers, who spend a lot of time on Twitter consuming news but don’t tweet often, are often identified as bots, as are accounts of non-English speakers who mistranslated a few words. In particular, avid fans who post frequently and in quick succession can appear as bots. And there will always be burner accounts owned by real people so they can anonymously engage in offensive behavior such as harassment. As a self-proclaimed free speech absolutist, Mr. Musk says he wants more freedoms on Twitter; in this context he seems to be fighting for less.
Mr. Musk has wondered if advertisers can know what they are getting for their money without understanding how many accounts are actually human. It’s a legitimate concern, but certainly not specific to Twitter. Mr. Kazemi described Twitter as one of the most open platforms with data, making it possible to study broadly academically over other platforms.
“Twitter is in bad shape,” Mr Kazemi said. “That puts it in the top 10% of all platforms.”
Twitter has historically not been afraid to look bad in order to get better. In 2019, Twitter shares plunged after the company said it would need further investments to clean up its user base. It also said it was moving from distributing monthly users to daily active users who could monetize, or those it thought could actually see ads.
In any case, instead of trying to compete with companies like Facebook and Instagram, Twitter has presented a low estimate of real users in an effort to show advertisers its value more accurately. Last month, the company publicly downgraded several quarters of user data, acknowledging that in some cases it had not linked multiple separate accounts.
SHARE YOUR THOUGHTS
Do you expect Elon Musk to go ahead with his Twitter takeover? Why or why not? Join the conversation below.
Meta Platforms said in its annual filing that duplicate accounts for its Facebook app may have represented about 11% of its global monthly active users in Q4 and that fake accounts may have represented about 5% of its global monthly user base. It called duplicate and false accounts “very difficult to measure” on its scale, noting “the actual numbers … could differ significantly from our estimates, possibly outside our estimated margins of error.”
The takeaway there: No major platform’s user numbers are useful for anything other than as random benchmarks used to benchmark against themselves.
Last week on Twitter, Mr. Kazemi revealed that Mr. Musk’s own Twitter account was flagged as a potential bot by Botometer, assigning a bot probability of 3.5 out of a potential 5.
Mr Kazemi said his work with online platforms to find out accurate user numbers was so futile that he gave up on his career. In his case, at least it was a legitimate excuse to walk.
write to Laura Forman at email@example.com
Copyright ©2022 Dow Jones & Company, Inc. All rights reserved. 87990cbe856818d5eddac44c7b1cdeb8