EXCLUSIVE: As the Trump administration works to keep TikTok legally available in the United States, the wildly popular app has suppressed content critical of President Donald Trump, according to a new report shared exclusively with Fox News.
TikTok maintains the report has reached a false conclusion, and that the researchers used terms subjected to additional safety measures because they’ve been associated with election misinformation or profanity.
The report, from the Network Contagion Research Institute (NCRI) at Rutgers University, contained findings that “highlight TikTok’s ability to act as a powerful influence tool, adaptable to partisan politics, but with no inherent incentive for transparency or accountability.”
CHINESE AI STARTUP DEEPSEEK FACING HACK, BLOCKS QUESTIONS ABOUT COMMUNIST PARTY TIES
“What you’re seeing is not sweeping policies around content moderation that can be battle tested by the public or by researchers,” said Adam Sohn, an NCRI board member. “TikTok seems to be just sort of picking and choosing their policies based on political expediency, and that’s a big concern.”
NCRI said it analyzed TikTok, X, and Instagram “to evaluate their handling of specific hashtags associated with the 2020 election controversy” and that researchers received a response that “explicitly indicated content suppression based on TikTok’s enforcement of its community standards.”
The group said terms such as “#RiggedElection,” “#VoterFraud,” “#StopTheSteal,” and “#StolenElection” returned no results on TikTok in the U.S. Researchers said that when they searched using software that swapped their domestic location for one overseas, those terms produced video results.
Screen grabs provided by NCRI show a Jan. 24 TikTok search for “#F***JoeBiden” that returned 37,000 results. A search the same day for “#F***Trump” returned none. Three days later, Fox News replicated the search and there were videos listed under both.
REPUBLICAN STATE AGS AWAIT TRUMP-BROKERED TIKTOK DEAL, REMAIN SKEPTICAL ON APP SAFETY
“The concern is that the Chinese Communist Party and Bytedance and TikTok itself can consistently tweak its algorithm to cover up its tracks,” Sohn said.
“Our policies and algorithms haven’t changed in the last week,” said a TikTok spokesperson.
The company maintains hashtags regarding the 2020 election controversies have promoted election misinformation, which is why they’ve been unavailable. TikTok contends that because the anti-Trump and anti-Biden search terms contain profanity, the app can limit those results. The company also says it’s experiencing technical issues as it’s trying to return its service to normal.
Last year, Congress passed a bipartisan law that would ban TikTok if its Chinese parent company, ByteDance, failed to sell the app by Jan. 19. Earlier this month, the U.S. Supreme Court upheld the law. ByteDance still owns TikTok, but Trump signed an executive order delaying the ban’s enforcement for 75 days while his administration tries to negotiate an agreement for the app to comply with the law and keep it operating in the U.S.
NCRI has issued several reports on TikTok, concluding its search algorithm produced results to construct a favorable view of China’s government. TikTok has denied that allegation, calling NCRI’s work “flawed” and “clearly engineered to reach a false, predetermined conclusion.” In its arguments against TikTok, the Justice Department under the Biden administration cited NCRI’s reports.
Cybersecurity experts told Fox that algorithms for apps like TikTok are held closely by their parent companies and can be difficult to evaluate.
“Doing sort of this community management of these vast social media platforms, especially TikTok, which is so popular, is a Herculean task,” said Theresa Payton, a cybersecurity expert and the White House Chief Information Officer in the George W. Bush administration. “It could be that as they were making tweaks to handle capacity, to be able to more closely evaluate things that could be perceived as election interference, things that are considered hate speech.”
Others note social media companies have sizable teams working with automated software to moderate content on their platforms.
“Someone interprets something as in terms of a violation [that] may not match with someone else – it all sort of has to add up to a pattern,” said Pete Pachal, the Founder of The Media Copilot, a newsletter on AI changing media and journalism. “In the report, they do a very good job of showing that this pattern of supposed repression … content not appearing in searches does tend to happen more in one direction, and that should arouse a certain amount of suspicion.”
Read the full article here