AG 1, AG 2, AG 3, INET, AG 4, AG 5, D6, SWS, RG1, MMCI   Info

AG Audience

English

30 Minutes

Short video platforms have become critical components of modern digital ecosystems, offering convenience while also introducing risks, such as illicit promotional content (IPC). Exploiting these platforms' dynamic content and wide reach, miscreants use sophisticated evasion techniques to bypass detection mechanisms, promoting their offerings behind the facade of legitimate content and delivering them to end-users.
To address this pressing issue, we designed an automated pipeline that combines natural language processing and video processing techniques to detect Short Video-Illicit Promotional Content (SV-IPC). Using a dataset collected from Chinese TikTok with 113,714 posts, we identified 513 instances of SV-IPC. Based on this, we conducted the first systematic study focused on short video platforms to uncover the ecosystem behind such activities, including adversarial strategies like text manipulation, frame insertion, and jargon usage to evade detection systems, while also quantifying the characteristics of miscreants. These findings shed light on the operational tactics of illicit promotion and highlight the need for increased detection strategies and platform policies to ensure social media platform integrity and user safety.

Ina Geisler

+49 681 9325 1802

--email hidden

Big Blue Button

passcode not visible

logged in users only