TikTok’s algorithm shows bias in political content, CCP-related coverage, studies find

Recent research reveals that TikTok’s recommendation algorithm exhibits significant biases in both its distribution of U.S. political content and its handling of Chinese government-related topics, raising fresh concerns about the platform’s influence on public opinion and electoral discourse.

Skewing Republican in 2024 election

A preprint study by researchers at New York University Abu Dhabi found that TikTok’s algorithm favored Republican-aligned content during the 2024 U.S. presidential election period. The research, which analyzed approximately 394,000 videos across three politically diverse states between April 30 and Nov. 11, 2024, discovered that GOP-seeded accounts received 11.8% more party-aligned recommendations compared to Democratic-seeded accounts.

The study revealed several key patterns:

Democratic-leaning accounts were exposed to roughly 7.5% more opposite-party content than their Republican counterparts
Negative partisanship content was particularly prevalent, with videos criticizing opposing parties being 1.78 times more likely to be recommended
Donald Trump’s official TikTok channel reached Democratic-conditioned accounts 26.9% of the time, while Kamala Harris’ videos were recommended to Republican-conditioned accounts only 15.3% of the time

Systematic bias in China content

A separate study published in Frontiers in Social Psychology investigated TikTok’s handling of content related to the Chinese government and found systematic differences in how the platform presents that content compared to other social media platforms.

The paper, a collaboration between researchers at Rutgers University and the Network Contagion Research Institute (NCRI), revealed that TikTok searches yielded significantly less content critical of the Chinese Communist Party (CCP) compared to Instagram and YouTube. For example, only 2.5% of search results for “Uyghur” on TikTok were coded as anti-CCP, compared to 50% on Instagram and 54% on YouTube.

The study also found a striking disconnect between user engagement and content distribution on TikTok:

Users engaged with anti-CCP content nearly four times more than with pro-CCP content.
Despite this engagement pattern, TikTok’s algorithm produced nearly three times as much pro-CCP content.
Higher TikTok usage correlated with more favorable views of China’s human rights record.

Broader implications

Both studies highlight concerns about TikTok’s potential influence on public opinion and democratic discourse. The platform, which boasts over 170 million users in the U.S. alone, has become a significant source of news, particularly among younger demographics.

The findings suggest that TikTok’s algorithm may be shaped by factors beyond typical commercial considerations. While the studies cannot definitively prove intentional manipulation, they reveal patterns that differ significantly from other major social media platforms.

Future research directions

The NYU Abu Dhabi researchers outline several key directions for future research: conducting longitudinal studies beyond election periods, combining automated experiments with real user data, analyzing visual content alongside transcripts and investigating misinformation spread specifically related to the 2024 elections. They also emphasize the importance of comparative studies across different social media platforms to better understand TikTok-specific effects.

Meanwhile, the Rutgers and NCRI researchers emphasize an urgent need for greater transparency in social media algorithms and call for developing robust methods to detect algorithmic manipulation. They argue that future research should prioritize ways to identify when platforms suppress information and undermine free expression, particularly in the context of authoritarian influence on democratic values.

This story is part of The Rebel Yellow Newsletter — a bold weekly newsletter from the creators of NextShark, reclaiming our stories and celebrating Asian American voices.

Subscribe free to join the movement. If you love what we’re building, consider becoming a paid member — your support helps us grow our team, investigate impactful stories, and uplift our community.

Subscribe here now!

Share this Article