Though current conversations around automated content moderation often focus on newer contexts for this technology such as harassment or misinformation, one other important use case is copyright detection systems such as YouTube's Content ID. This context for content removals online has some unique properties such as its relationship to legal requirements (Section 512 of the Digital Millennium Copyright Act that governs copyright notice-and-takedown), and provides a longstanding example of the intersection between policy, online platform design, and algorithms. Based on qualitative analysis of survey data, this paper examines the impact of DMCA takedowns and automated copyright-related content removals on a community of content creators who primarily share transformative works, remixed content that makes use of copyrighted material. Our data reveals patterns of chilling effects related to prior experiences and perceptions of policies and process. Based on these findings we make recommendations for platform policy and design, education, and advocacy, and discuss the implications of our findings for current and proposed laws.
