Researchers from Spring AI claim TikTok’s recommendation algorithm is favoring content supporting Zohran Mamdani, a socialist candidate for New York City mayor, while suppressing posts backing former governor Andrew Cuomo. The findings, disclosed by the company’s chief technology officer, Yehonatan Dodeles, suggest the platform’s algorithm systematically amplifies pro-Mamdani and anti-Cuomo content during the mayoral race.
The study identified “Excessive Publicity” videos—posts receiving disproportionate exposure compared to organic reach—which accounted for 17% of all videos analyzed. However, political content showed a stark disparity: 55% of videos in this category were boosted, with most promoting Mamdani or criticizing Cuomo. Pro-Cuomo content was reportedly suppressed below baseline levels.
Dodeles warned that algorithmic steering could shape not only user visibility but also public perception, raising concerns about social media’s role in influencing political outcomes. The research highlights broader debates over platform accountability and the potential for digital tools to sway democratic processes.
The findings come amid heightened scrutiny of foreign influence in U.S. elections, including allegations of Chinese Communist Party efforts to interfere with mail-in ballots during the 2020 cycle. However, the study focuses on TikTok’s internal mechanisms rather than external actors.
Spring AI emphasized that its research is preliminary and called for further investigation into the implications of algorithmic bias in political discourse.
