An unverified study, conducted by a former Israeli intelligence operative, suggests that TikTok's algorithm may be manipulated to influence political opinions, particularly concerning Donald Trump's policies towards Iran. The findings, shared with former UK Prime Minister Boris Johnson, propose that content critical of Trump's actions regarding Iran was disproportionately recommended to users.
The experiment involved creating 59 bots with TikTok accounts that passively scrolled through the platform's content for approximately one month around March of the current year. Researchers analyzed the publicly available data of over 6,500 videos these bots encountered. While the algorithm's behavior for most content could be predicted based on engagement metrics like views, likes, and shares, a specific category of videos—those mocking or criticizing Trump's stance on Iran—reportedly appeared 21% more frequently than expected.
The study authors acknowledge the possibility of a statistical anomaly but estimate the chance of such a result being coincidental at one in 30 million. This alleged over-recommendation of anti-Trump content has led to speculation that the Chinese owners of TikTok might be using the platform to shape Western political sentiment, potentially fostering skepticism towards Trump and Israel while promoting a pro-Iran and, by extension, pro-China stance.
However, the claims face scrutiny. Critics point to the provenance of the study and the potential motives of the Israeli intelligence community, given global disagreements on the U.S. approach to Iran. Furthermore, the experiment's methodology may not fully capture the complexity of TikTok's algorithm, which could account for user engagement factors not detected by the bots, such as watch time or further research into video creators. TikTok's algorithm may also utilize data not accessible to the bots in the experiment.
Johnson highlighted the broader implications of social media algorithms, noting that platforms like TikTok, as well as American-based ones such as Google, X, Instagram, and Facebook, operate largely outside of the UK's regulatory oversight. He expressed concern that these algorithms, designed to cater to user preferences, could be inadvertently or intentionally reinforcing societal divisions and polarizing political discourse. The lack of transparency and accountability for these powerful algorithms, which significantly influence public opinion, especially among younger demographics, is a key concern.
"The algorithm watches you. The algorithm knows you. The algorithm loves to make money out of you," Johnson stated. "The algorithm can clearly manipulate the way you think. It’s time we knew more about the algorithm and the people behind it."