According to Cameron Hickey, project director at the Algorithm Transparency Institute, TikTok is intrinsically more difficult to censor than many other social media platforms. The brevity of videos and the fact that many videos can include audio, visual, and text elements make human discrimination all the more essential when deciding if something is infringing. the rules of the platform or not. Even advanced artificial intelligence tools, like using speech-to-text to quickly identify problem words, are harder “when the sound you’re processing also has music.” behind it,” said Hickey. “The default mode for content creators on TikTok is to also embed music.”
That becomes even more difficult in languages other than English.
“What we know in general is that platforms work best at solving content that matters where they are based or within the language their creators speak,” says Hickey. “And there are more people making worse things than the people at these companies trying to get rid of the bad stuff.”
Much of the misinformation Madung found was “aggregate content”, videos made to look like they were possibly from an old news broadcast, or they used screenshots that appeared to be from legitimate news agencies.
“Since 2017, we have noticed that there is a growing trend at the time that is consistent with the identity of mainstream media brands,” says Madung. “We’re seeing widespread use of this tactic across the platform and it seems to be working very well.”
Madung also spoke with former TikTok content moderator Gadear Ayed to gain insight into the company’s broader moderation efforts. Although Ayed does not moderate TikToks from Kenya, she told Madung that she is often asked to moderate content in a language or context she is not familiar with, and would have no context to know if a method is being used. whether the media has been manipulated or not.
“Moderators are asked to moderate videos with language and context different from what they normally understand,” Ayed told Madung. “For example, I once had to censor videos in Hebrew even though I didn’t know the language or the context. All I can rely on is a visual representation of what I can see but whatever is written I cannot censor.”
A TikTok spokesperson told WIRED that the company prohibits election misinformation and promotes violence and is “committed to protecting the integrity of [its] platform and has a dedicated team working to protect TikTok during elections in Kenya. The spokesperson also said it works with fact-checking organizations, including Agence France-Presse in Kenya, and plans to roll out features to connect “its community with authoritative information.” about elections in Kenya in our app”.
But even if TikTok does remove the offending content, Hickey says that may not be enough. “One person can remix, duet, re-share another person’s content,” says Hickey. That means even if the original video is deleted, other versions can still exist undetected. TikTok videos can also be downloaded and shared on other platforms, like Facebook and Twitter, which is how Madung first encountered some of them.
Some of the videos flagged in the Mozilla Foundation report have since been removed, but TikTok did not respond to questions about whether it removed other videos or whether the videos themselves were part of the report. coordinated effort or not.
But Madung doubts that they could be. “Some of the most egregious hashtags are the ones I would find researching coordinated campaigns on Twitter, and then I would think, what if I looked this up on TikTok?”
