The discussion regarding racial unfairness on TikTok was caused by the hate speech detection tool.
ByShehryar Makhdoom | Published date:
To rectify a mistake that has led BLM content to be flagged by its algorithms, TikTok claims it has updated its software. However, some Black producers are still concerned.
"I'm fucked off because of that. We're exhausted," prominent Black influencer Ziggi Tyler stated in a new TikTok viral video. "We reserve the right to remove all black-related content," he said in the video.
During the editing of his bio on the Creator Marketplace app, Tyler expressed his frustration over TikTok, which links popular account holders to advertisers who pay for items or services to promote.
Tyler noted that the app marked its contents as "inappropriate" when he entered keywords regarding Black content in his Marketplace-based organic designer, such as "Black Lives Matter" or "Black achievement."
However, when he entered the phrases "white supremacy" or "white success," he was not warned.
Tyler and many of his fans perceived the episode as part of a bigger trend of how Black content on social media is controlled. Some supporters of the charge called for the public to boycott the app, and others tagged TikTok's corporate account to elicit a response.
The original video describing the incident, which Tyler uploaded, has got over 1.2 million views and roughly 25,000 comments.
A representative from TikTok informed Recode that the problem stemmed from an error in its hate speech detection systems and that it is actively working to fix the problem.
The problem does not show evidence of racial bias, according to the spokesman. According to a spokesman, TikTok standards do not limit posting on Black Lives Matter.
TikTok has stated to Recode indicating that it wrongly flags terms like "Black Lives Matter" as hate speech because its hate speech detector is activated by a phrase containing "Black" and "audience."
A statement from the company said: "The inaccurate spelling rules and options that we have built into our TikTok Creator Marketplace protections, which flag words typically associated with hate speech, were set erroneously, flagging comments without regard to the words' order.
"We acknowledge and apologize for the frustration this caused, and our staff is working diligently to correct this important blunder. Black Lives Matter is in full compliance with our guidelines, and their video has been viewed over 27 billion times on our platform.
TikTok claims to have contacted Tyler directly and received no response.
An AI-powered algorithm has been shown to identify tweets authored by African Americans 1.5 times more as "offensive" than tweets written by other Americans.
Findings like these have sparked a debate regarding the benefits and drawbacks of using algorithms — especially developing AI models — to automatically discover and censor social media content.
However, even though these algorithmic models are ineffective in the past, major social media companies such as TikTok, Google, Facebook, and Twitter continue to use them as a key component of their fast-growing hate speech detection systems.
They claim they require a less time-consuming method of keeping up with the ever-increasing volume of stuff on the internet.