Ann Arbor Times

A recent study from the University of Michigan finds that users who actively search for conspiracy-related content expressing anti-establishment sentiment can regularly find such material on TikTok and YouTube. The research, titled "Anti-establishment sentiment on TikTok: Implications for understanding influence(rs) and expertise on social media," highlights how distrust in institutions—such as the media or health care—is not prevalent in most content categories but is concentrated within specific themes.

According to the study, less than 1% of all content on a typical For You Page—the algorithm-driven feed that tailors videos to user interests—contains anti-establishment sentiment. This suggests that casual users are unlikely to encounter this type of messaging frequently.

Researchers collected 26,783 video posts and analyzed 206,350 comments over two years beginning January 1, 2022. The analysis focused on three main themes: finance, wellness, and conspiracy theories. They found that about 45% of conspiracy-themed posts included anti-establishment sentiment, compared to just 4% in finance and 1% in wellness topics.

Tianliang Xu, doctoral student at the U-M School of Information and lead author of the study, stated: “We see that people who take the time to comment on anti-establishment sentiment posts tend to agree with the message in the video, which suggests that those who are drawn to this content are likely to be recommended more of it.”

The study also examined which institutions were targeted by these sentiments. Researchers hand coded a sample of posts predicted to contain anti-establishment messages and observed that vague entities like “the government,” “them,” or “big pharma” were mentioned more often than specific organizations. However, some institutions such as NASA were cited frequently.

While everyday content rarely contains these messages, when anti-establishment views do appear in finance or wellness posts they tend to generate higher engagement through comments. Most commenters agreed with the original post even if it contained controversial claims.

Sabina Tomkins, assistant professor of information and co-author of the study said: “While we expect that most platforms will not show anti-establishment sentiment content to an average user often, they may vary in how their algorithms show such content to interested users.”

Ariel Hasell, associate professor of communication and media at U-M, also contributed as a co-author.

The paper received partial funding from the U-M Presidential Award for Understanding Democracy and will be presented at the International AAAI Conference on Web and Social Media next spring in Los Angeles.