20 July 2021
Twitter relaunched its public verification program this year in May and allowed anyone to apply for it. Many applied to this program and did get the verification after fulfilling the revamped criteria, however, the program also became a host for several fake accounts. A data scientist observed that Twitter has also verified at least six fake accounts. Following the report, Twitter also condemned it and admitted that it “mistakenly” verified the accounts.
The discrepancy was spotted by Conspirador Norteno, who shared the findings via a tweet thread and Daily Dot publication reported it later. Norteno said at least six accounts have got verified by Twitter’s newest verification program. The scientist observed that all these accounts have 0 tweets and 0 followers and yet they have a blue tick right beside their username. All these accounts roughly had 1000 followers, all of which were created between June 19th and June 20th, 2021.
The sleuth was able to point out that no profile image was able to depict the account holder as images appear to be stolen. These images appear to have been generated using GAN (Generative Adversarial Network), an AI technique.
Later, Mikael Thalen from Daily Dot reported that Twitter has confirmed to him about the mishap and said it has now suspended those accounts and removed the verified badge.
“It mistakenly approved the verification applications of these fake accounts. We have now permanently suspended the accounts in question, and removed their verified badge under our platform manipulation and spam policy.”
As TheVerge points out, there appears to be some issue with Twitter’s newest verification process as it is not able to detect and weed out the accounts that are not authentic. In its revamped rules of verification, Twitter has mentioned that you’re worthy of receiving the blue badge if your account is “authentic, notable, and active.” But, clearly, the program couldn't notice all these accounts that have failed the policy.
This is sad to note because Twitter had mentioned during the release of this new program that this “new verification policy will lay the foundation for future improvements by defining what verification means.” But, if the foundation is weak, the possible future improvements might not sustain well. The newest process should have flagged these fake accounts in the early detection. It is daunting that Twitter took action only after a third-party researcher reported them.