The Social Impact and Cross-cultural Analysis of Short-Video Recommendation Algorithms on TikTok

Abstract

: As short-video platforms become deeply embedded in global digital life, recommendation algorithms have emerged as a core force shaping contemporary information dissemination and user cognition. From the perspective of communication studies, this paper adopts the narrative literature review method to systematically sort out the mechanism, social impact and cultural differences of short-video algorithms represented by TikTok/Douyin. The study finds that: firstly, the relationship between algorithms and users is not one-way control, but a two-way dynamic feedback loop of "data input - algorithm prediction - behavior adaptation"; secondly, while promoting the democratization of content production, economic empowerment and knowledge popularization, recommendation algorithms have also triggered profound social risks such as digital addiction, political polarization and algorithmic discrimination; thirdly, the large-scale application of generative AI is extending the power of algorithms from "content distribution" upstream to "content production", exacerbating the governance difficulty of copyright infringement and the proliferation of misinformation. Finally, a comparison based on Hofstede's cultural dimension theory shows that the algorithmic logic of Douyin and TikTok reflects deep cultural differences between the East and the West in collectivism/individualism and power distance. This paper argues that short-video algorithms are not value-neutral technical tools. Future algorithm governance needs to strike a balance in transparency, embedded pluralistic values and international collaboration, so as to rebuild users' digital subjectivity under the "algorithmic gaze".