“I and other people who have tried to help people have failed,” says Benavidez. “And when we reached out to those who said they were still on Twitter, we didn’t respond.”
Although researchers can go through Twitter, responses are slow—sometimes taking more than a day. Jesse Littlewood, vice president of campaigns at the non-profit Common Cause, says he realized that when his organization reported tweets that clearly violated Twitter’s policies, the posts could now not be removed.
The number of things that users and guardians would like to share on Twitter is bound to increase. Many of the employees and contractors who have resigned in recent weeks worked in areas such as trust and security, policy, and public integrity, all of which helped to keep the public from disinformation and hate speech.
Melissa Ingle was the chief data scientist on Twitter’s social media team until she was laid off along with 4,400 other contractors on November 12. She wrote and managed the algorithms used to identify and remove political errors on Twitter—latest , which means that the election in the US. and Brazil. Of the 30 people on his team, only 10 remain, and most of the human content moderators, who monitor tweets and flag content that violates Twitter’s policies, have also been fired. “Machine learning needs constant support, constant maintenance,” he says. “We have to constantly change what we want because political issues are always changing.”
Although Ingle’s work did not involve interacting with activists or researchers, he says members of Twitter’s political community did. In some cases, information from outside groups helped inform Ingle’s findings and his team trained algorithms to identify them. Now they are worried that with so many workers and contractors out of a job, there won’t be enough people to make sure the program stays accurate.
“With the algorithms not updated and the human moderators gone, there aren’t enough people to steer the ship,” says Ingle. “My concern is that these filters are becoming more and more complex, and more things are going to come in because the algorithms are getting weaker over time. And there’s no one to catch things falling through the cracks.”
Within a day of Musk taking ownership of Twitter, Ingle says, internal data showed that the number of abusive tweets reported by users had increased by 50 percent. The initial spike has slowed slightly, he says, but abuse reports have remained about 40 percent or higher than normal volume before the takeover.
Rebekah Tromble, director of the Institute for Data, Democracy & Politics at George Washington University, also expects to see Twitter’s defenses against censorship weaken. “Twitter has been struggling with this, but a number of talented teams have stepped up to the plate in recent months. Those teams are completely gone.”
Such concerns are confirmed by an archivist who was a contractor for Twitter until 2020. The contractor, speaking on condition of anonymity to avoid the consequences of his current employer, says that all his former colleagues who worked with him in the same work have been fired. He hoped that the tower would not be a good place to live. “It’s going to be bad,” he says. “I’ve quickly explored the worst parts of Twitter – the most racist, the most dangerous, the most degrading of the platform. That’s what will be expanded.”