Skip to main content

Twitter admits it verified several fake accounts

Twitter admits it verified several fake accounts

/

Part of a 1212 account-strong botnet

Share this story

Illustration by Alex Castro / The Verge

Twitter says it has permanently suspended a “small number” of fake accounts it mistakenly verified just weeks after re-launching its public verification program, the Daily Dot reports. The blunder came to light after data scientist Conspirador Norteño discovered six verified accounts which had been created recently on June 16th. None of them had posted a single tweet, and two used what appeared to be stock photographs for profile pictures. 

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

The incident suggests that Twitter’s verification process is having problems, and not catching the kinds of obviously inauthentic accounts that shouldn’t be worthy of the platform’s coveted blue badge. Twitter recently relaunched public verification applications with a revamped set of eligibility criteria based around the idea that an account should be “authentic, notable, and active” to be worthy of verification. Clearly, the accounts identified were none of these.

As Norteño explains in a Twitter thread, the six accounts had 976 suspicious followers in common — all created between June 19th and June 20th — and large numbers of these were using AI-generated profile pictures. In total, Norteño says they were part of a botnet consisting of at least 1212 accounts.

As of this writing, Twitter has suspended five out of the six verified accounts, while the sixth appears to have deactivated its own profile. The “majority” of the supporting botnet has also been taken down, Norteño says. But the incident raises questions about how the accounts were able to get verified in the first place, and why Twitter’s processes didn’t flag them before they were discovered by a third-party researcher.