Twitter Bots, the Untold Story

Social media isn't just becoming more fake by accident

Hey Guys,

When social media companies report to shareholders of how many million DAU, MAUs and users they have, are they reporting real people?

As users, we know fake accounts, bots and deepfake accounts are everywhere, but there’s little real data and transparency on the issue. The Twitter bot controversy sheds a lot of light on this.

Silicon Valley has been Lying about its Subscriber Numbers for Years

Twitter estimated in a filing earlier this month that fewer than 5% of its monetizable daily active users during the first quarter were fake or spam accounts. Everybody including recent research knows it’s far higher than that. And it’s a major black eye on Silicon Valley’s social media, a lot of the supposed users they report, aren’t real people at all.

Musk, who has hinted that the prevalence of fake bots may pull out of his $44 billion acquisition of Twitter, has a total of 93.8 million followers on the social media platform. So if Twitter is right and less than 5% of those are fake, it wouldn’t be such an issue.

Twitter has a Bot and Fake account Crisis

Recently, SparkToro and Followerwonk audited Musk’s Twitter followers.

Both companies said they used a system of 17 warning signs. They also deployed an algorithm that scanned through 35,000 fake accounts purchased by SparkToro.

Nearly one out of every four of Elon Musk’s Twitter followers are either spam or bot accounts, according to researchers.

According to two research groups reported by Business Insider, more than 23.42% of them have either fake or spam accounts.

I can believe it because it mirrors my experience on LinkedIn and Facebook in recent years. Easily between twenty to thirty percent of accounts you come across are likely fake. So you have to take their “user numbers” with a grain of salt.

Then there is another huge factor, inactive accounts. For instance, the two research firms also estimate that more than 70% of Musk’s followers are accounts that haven’t tweeted in 90 days.

According to the researchers, these inactive users were unlikely to be “authentic” or “active users who saw their tweets.” Powerful people use “followers” as a kind of social proof to their voice, but a lot of it is fake. It’s purely fictitious.

Musk claims that as many as 20% of Twitter’s users could be bots. According to Israeli tech firm Cyabra, bots could make up some 13.7% of Twitter's 229 million total accounts.

Malicious actors can also use bots to scam people out of cryptocurrency or spread disinformation. Countries like Russia have been accused of using bots to spread false information about vaccines and U.S. elections. And unfortunately, these malicious bots have ever-evolving ways of evading detection.

LinkedIn’s China fake account problem was so bad, they literally had to close LinkedIn China, although the reasons given were not necessarily about the espionage and hacking they were experiencing.

Musk and the Twitter Bot Controversy

Musk has been accused of exploiting bot situation to get out of buying Twitter. After all, Twitter has reported that bots make up less than 5% of its total users every quarter since at least 2014. However this 5% figure is not factual when compared to recent research on the problem. Twitter’s under-reporting of the true number is making apparently Elon Musk think twice about his bid. Even this however, is likely to be a legal stunt rather than a real concern of Elon Musk.

Company CEO Parag Agarwal took to Twitter on Monday in an attempt to refute Musk’s claim that the social media site allowed spam and bot accounts to run amok. Spam “harms the experience of real people on Twitter,” he wrote, and the company is “strongly encouraged to detect and remove as much spam as possible every single day.” However if this is a fight, LinkedIn, Facebook and Twitter are mostly failing to take the problem of bots, fake accounts, inactive accounts and deepfakes.

Web2 is Leaning to “Synthetic Engagement” Tactics to Boost Numbers

From crypto projects on Twitter to the lack of new users who are real, you can see the exodus from Web3 platforms like Facebook, LinkedIn and Twitter in real-time. Now new deepfake technology makes it easy for people to create fake accounts, bots and synthetic content at scale that could be a mix of foreign actors, and increasingly likely, the companies themselves. It is however a deceptive factor in view of shareholders.

That Twitter is saying 5% while the reality is likely so much higher is really concerning for the validity and credibility of these platforms. I can report ten fake accounts on LinkedIn per day, but would it make a difference? It isn’t my job.

In an increasingly synthetic internet, it’s not my job to figure out what is real or not. If I don’t trust your platform, I will likely leave or at least be less active eventually.

Musk estimates that around 20% of the accounts on Twitter are fake or spam accounts and he’s concerned that the number could be even higher. That’s 4x what Twitter admit publically. I think we can assume on most social media platforms of scale, the real number is closer to 25%. So if Facebook is boasting 2 Billion users across its suite of apps, what is the real number?

Back in 2017, a study suggested 48 million accounts on Twitter, weren’t even real people. In all fairness, that’s a lot of Twitter’s total user base in 2017. The study suggested 15% were fake and I think with deepfake tech getting better it could be a lot worse today. According to the 2017 research from the University of Southern California and Indiana University, up to 15 percent of Twitter accounts are in fact bots rather than people.

The take home here I think is Twitter is not alone in having a good chunk of its “engagement” as actually “fake”. You can call them bots, synthetic, inactive accounts or whatever you like, they are just not real people or real users. Silicon Valley has had this issue for years, and it makes their monthly active user accounts especially fake. Linkedin like to say it has 800 million users, but since LinkedIn China was axed, it’s been using the same number. It’s beyond redemption.

In the end we’ll have to look at Ad-revenue and App downloads to make up our minds on how popular platforms and apps really are, but with TikTok taking over, who really knows if those numbers are manipulated as well.

Silicon Valley has a lot more problems than just bots, but a major exodus from their formative apps of the internet. Including false measures of engagement that can be manufactured. What’s the incentive for LinkedIn not to manufacture “false engagement” signals for its Microsoft managers? I see it every day. It means social media (outside of some Reddit channels) is becoming increasingly fake, synthetic and manipulated.

Personal

With my Mom in the hospital with terminal Cancer in her last few days, I may be far less active on Substack, LinkedIn and in my usual writing output in May, 2022. Thanks for your understanding.

Thanks for reading!