Analysing Tiktok's "Others Searched For" Feature
Our study examines how TikTok's search suggestions influence political discourse among young German users ahead of 2024 elections.
Project overview
Conducted with interface's TikTok Audit Team, this research investigates TikTok's 'Others Searched For' feature to understand whether TikTok's algorithm promotes misleading or sensational content through search suggestions in the context of the 2024 German elections.
Findings:
- 67% of young German TikTok users (18-24) regularly use the search function, frequently encountering TikTok's search suggestions
- Search suggestions are often phrased as clickbait, seemingly selected for engagement rather than content relevance
- We collected 254 search suggestions across 9 political parties over 4 weeks leading to the election
- 25% of suggestions redirect to other parties, with AfD being the most frequently referenced
- Within redirecting search suggestions, AfD dominates—appearing in suggestions for all parties except Die Linke and FDP
- 8% of suggestions contain dog whistles potentially referring to right-wing narratives (e.g., 'califat in Germany', 'putin warns turkish people in Germany')
Featured On
Related Projects
Searching for ModerationFollowing ChatGPT Search's launch just before the US election conclusion, we tested its content moderation effectiveness. Our research found alarming gaps: the search feature provides links to EU-banned Russian state media outlets, misattributes legitimate reporting to Kremlin-affiliated sources, and applies inconsistent moderation to election-related content.29-11-2024
TikTok’s Role in Youth Mental HealthTeaming up with the Algorithmic Transparency Initiative and Amnesty International, we conducted a two-part audit on TikTok's content recommendation system. Our findings reveal a stark and troubling reality: TikTok's algorithms could potentially lead young users down dangerous paths, exposing them to content that glorifies self-harm and suicidal thoughts.07-11-2023
Searching for Moderation
Following ChatGPT Search's launch just before the US election conclusion, we tested its content moderation effectiveness. Our research found alarming gaps: the search feature provides links to EU-banned Russian state media outlets, misattributes legitimate reporting to Kremlin-affiliated sources, and applies inconsistent moderation to election-related content.
29-11-2024
TikTok’s Role in Youth Mental Health
Teaming up with the Algorithmic Transparency Initiative and Amnesty International, we conducted a two-part audit on TikTok's content recommendation system. Our findings reveal a stark and troubling reality: TikTok's algorithms could potentially lead young users down dangerous paths, exposing them to content that glorifies self-harm and suicidal thoughts.
07-11-2023
AI Forensics is proud to be trusted by
