Twitter’s Verification Process Remains Problematic, with Several Bot Profiles Getting Blue Checkmarks

Have you recently applied for Twitter verification, and been subsequently rejected by the platform’s assessment team?

This might help to soften the blow – despite launching a new and improved verification process back in May, designed to help signify notable identities in several defined categories, researchers have found that even Twitter’s new system is still flawed, potentially in a big way, with several bot profiles, using entirely fake identities, being granted the prestigious blue tick.

A data researcher going by the name of Conspirador Norteño on the platform has identified six bot accounts, each created less than a month ago, that have been approved for verification, somehow making their way through Twitter’s new, more rigorous testing process.

As explained by Norteño:

“These six newly-created verified accounts have 977 followers in common. One is @verified (which follows all blue-check verified accounts). The other 976 were all created on June 19th or June 20th, 2021, and all follow the same 190 accounts. […] Very few of the accounts in this network have tweeted. The majority of the tweet content is spam in Korean sent via automation service dlvr(dot)it promoting a website.”

Norteño further explains that many of the profile images these accounts have used are either stock photos or AI-generated depictions, meaning that they’re almost definitely fake accounts – or at best, they’re not honestly representing any person or business. 

Shortly after Norteño’s tweet started getting attention, Twitter deleted the accounts, while also providing this confirmation (to Daily Dot) that they had indeed, somehow, been verified, despite the many red flags:

“We mistakenly approved the verification applications of a small number of inauthentic (fake) accounts. We have now permanently suspended the accounts in question, and removed their verified badge, under our platform manipulation and spam policy.”

So how does that happen? How does Twitter, with its updated, more vigorous, more strenuous verification process, which is designed to ensure that the wrong people stop getting blue ticks, end up approving members of a bot network, that was likely looking to use the platform to spread harmful spam and/or disinformation? 

Without internal insight into the process, we don’t know, but there are a few suggestions as to what the flaw, or flaws, may be.

One theory is that Twitter’s new verification process is entirely automated, or at least, automated to the degree that a human at the end of the chain can feel comfortable that various checks have been made by Twitter’s bots to get a profile through to the final approval stage.

If that’s the case, and spammers have worked out what, exactly, the system is looking for, they may well be able to get at least some of their accounts approved by getting as many of them as possible through the automated checking process.

At some point, with so many of these profiles coming through, the workload stacks up, and the human/s making the final call have to start ticking them off to clear the backlog. If all of these profiles have been pre-passed by automated checks, maybe Twitter starts approving the wrong profiles, and you end up with a situation where the worst actors on the platform now, effectively, have official endorsement, boosting their credibility, and potential for harm.

That’s a disastrous outcome for Twitter’s trumpeted return of verification applications, after almost five years in suspension.

In fact, given these new findings, it’s possible that Twitter’s updated verification system is actually worse than it had been before. If your application has been rejected, it could actually have come down to the most binary, basic error in the process, with the automated system misreading the information you’ve supplied – while scammers who’ve worked it out can get almost any profile approved, bringing the whole process into question.

What good is having a verification system if you can’t trust what that blue checkmark even means? And these are only six examples – how many others are there that have been approved simply because they met Twitter’s flawed requirement system?

To be fair, Twitter says that humans are assessing verification applications, and it’s not entirely automated.

In a recent Twitter Spaces session, members of the platform’s product team answered some of the common questions about verification, and why people are being rejected.

As noted in the session (via Financial Express):

“The people who review applications are undergoing rigorous training to ensure that we only verify eligible people, per our policy.”

You would think that ‘eligible people’ would need to represent real entities as a basic measure, but evidently not, and if Twitter is in fact using human review for all applications, how could it have possibly approved these bot profiles when even the most basic digging reveals some pretty significant cracks in their facade?

No matter how you look at it, it’s not a great endorsement for Twitter’s process. And while it could well detect and fix the errors, and weed out any profiles that have been mistakenly approved, it’s a pretty significant flaw, which suggests that Twitter hasn’t exactly worked out its verification system just yet.

Basically, if you were rejected, it’s probably not you, it’s Twitter’s broken system that made a mistake. 

At least you can hold onto that as a comfort for your dented ego in these trying times.