I turn my attention to Dublin, where I find Fabian (not his real name), a former moderator who worked for Twitter through the outsourcing company CPL until earlier this year. Unlike another moderator I spoke to who worked for Twitter, he recalls seeing accounts that the company had labeled “suspicious” internally. He doesn’t know who – or what – tagged those tweets. “I think the system used to work,” he told me. “Or maybe there’s a team dedicated to it.” Talking to him I’m sure someone, somewhere, is moderating Twitter bots and this effort is new and only started in recent years.
Another clue comes in late August, while I’m on summer vacation. Peiter Zatko (or Mudge), the former head of Twitter’s security, has decided to become a whistleblower and file a report commissioned by an independent company. to the US Congress. The document names the internal Twitter teams responsible for spam and other attempts to “manipulate” the platform at the time — one called Site Integrity, the other Health and Twitter Services (two teams Twitter has since merged). Another line in the report stands out. It says, “Content moderation is outsourced to suppliers, most of whom are located in Manila.”
Twitter has long used outsourcing companies to hire people in the Philippines to remove material about violence and sexual abuse from the site. But can they also moderate spam? In the industry there are the big, recognizable names such as Accenture and Cognizant. But there are also the lesser-known companies, such as Texas-based TaskUs. Finally, I come across a company I’ve never heard of: a New Jersey-based company called Innodata. And for the first time I hear the job description ‘spam moderator’.
I speak to an Innodata employee who confirms that the company moderates spam for Twitter even though he worked on a different team. Another says he has been involved in “categorizing” fake accounts, some of which pose as famous sports teams. Both ask that their names and locations not be published for fear of losing their jobs. According to a recent job posting, Innodata has approximately 4,000 employees in Canada, Germany, India, Israel, the Philippines, Sri Lanka, the United States and the United Kingdom.
By specifically looking for moderators at Innodata, I finally find John, the employee who shares the photo of the woman in the swimsuit. He explains that there are 33 full-time employees who moderate spam for Twitter and more than 50 freelancers. He thinks Innodata didn’t start moderating spam until March 2021.
Every day, John says he checks up to 600 Twitter posts and accounts in a third-party app called Appen, before marking them as “spam” or “safe.” (Appen is an Australian company that uses a global workforce to train artificial intelligence used by major tech companies.) The majority of John’s team is based in India or the Philippines, he says. He believes the tweets he sends are selected by artificial intelligence trained to look for Twitter spam before being forwarded to a team of human moderators.