Twitter’s new verification system has a serious flaw


Twitter’s new verification system has a serious flaw

Twitter is preparing to relaunch its verification system as early as Friday, but Elon Musk’s second stab at pay-for-play blue checks is unlikely to solve the platform’s growing problem with misinformation.

The billionaire first made coveted blue checks available last month to anyone willing to pay $7.99 per month, no proof of identity needed — quickly reversing after Twitter accounts impersonating brands shaved millions of dollars of companies stock prices and cost Twitter millions in paused ad spending.

Twitter is now set to introduce a series of different colored checks to identify governments and brands, while still allowing individuals to purchase blue checks. Musk has said each person will need to be manually verified. But experts are skeptical that Twitter’s greatly reduced staff will be able to suss out spammers, bots or impersonators. At the same, they see those bad actors as the most likely customers for the program, for whom the new price — $8 — is a low one to pay for a veneer of legitimacy.

“Having different colored verified checks for companies and government offices is all well and good, but still does not solve the problem of check marks for individuals,” said Yael Eisenstat, vice president of the Anti-Defamation League’s Center for Technology and Society. “Of course, you can verify the government of Australia or verify a celebrity like LeBron James. But what about regular people? How will he verify legitimacy for individuals?”


I’ll take the check, please

Musk has argued that his changes to the verification system will democratize it — and help solve the site’s bot, spam and misinformation problem. For years, Twitter awarded blue checks to a mix of government, media, celebrity and academic users without clear processes for interested parties to seek verification.

But tech industry observers are skeptical that Musk’s second stab at a pay-for-play verification system will fulfill the latter goal.

David Karpf, an associate professor in the School of Media and Public Affairs at George Washington University, said that the system, in fact, encourages scammers and others to join, as it gives them a pretext of authenticity that could give them a financial advantage that goes well beyond just $8.

“The stupid thing is the people who are most likely to pay for a blue check mark are the people who are going to see financial advantage,” said Karpf. “Crypto scammers are, I think, the only people who signed up for this check mark [during Musk’s first attempt], because it makes their scam marginally more effective. The fundamental theory was just garbage in ways that were really pretty obvious.”

Supporters of the fringe QAnon conspiracy movement also bought in during that first round of pay-for-play.


The Patriot Voice Reports, a QAnon conspiracy theory Twitter account disguised as a news site, has a verified check mark and all of around 800 followers. Created in October 2022, the same month that Musk bought Twitter, the account links back to the Patriot Voice website, run by John Sabal, also known as “QAnon John”

“Let’s remember that Twitter check marks portray an element of ‘credibility’ or ‘endorsement’ from the company itself,” said Eisenstat. “There’s a reason why QAnon accounts have paid for check marks already: They see it as a marker of legitimacy and a tacit endorsement of their conspiracy theories.”

About 140,000 people signed up for the new Twitter Blue last month, with the average account having around 500 followers.

That revenue is not nearly enough to make up for massive halts in ad buys by advertisers as they warily see how Musk proceeds. Grid reported previously that advertisers are asking about everything from hate speech to how Twitter stacked up next to its competitors when it came to online safety. The team that advertisers kept in touch with was also decimated by layoffs and resignations.

Those layoffs and resignations, which have reportedly reduced Twitter’s total headcount by more than half, will also make monitoring the verification system incredibly difficult — especially if that verification is manual, as Musk has promised.

“The idea you’re going to do it manually, while also getting rid of like 80 percent of your staff and 80 percent of your contractors — is David Sacks handling this or Jason Calacanis handling this?” said Karpf, referring to two close Musk advisers. “You don’t have enough bodies to manually handle all of this checking, and you definitely can’t throw machine learning at it. But he’s definitely going to try, because I don’t think he understands what machine learning does and does not do well.”

Until the fundamental misalignments around what the purpose of verification is, it’s hard to see even this iteration going well.

“History indicates that bad faith actors will find a way to exploit these systems, which is why carefully considered content and user policies matter,” said Eisenstat.

Thanks to Alicia Benjamin for copy editing this article.

  • Benjamin Powers
    Benjamin Powers

    Technology Reporter

    Benjamin Powers is a technology reporter for Grid where he explores the interconnection of technology and privacy within major stories.