Analyzing Copyright Infringement Liability in Web Scraping Cases—A Critical Examination of Recent Global Trends.

Abstract

Using copyright information from protected literary or artistic copyrighted works is on the rise as suggested by an influx of global cases on this subject matter. For instance, in Anderson v. Stability AI, et al. the plaintiffs alleged before the United States District Court of Northern California, San Francisco Division, that JPEG files generated by the defendants’ Stable Diffusion and Mid Journey Product AI have been derived from the protected artwork of the plaintiffs. But the claims were set aside by the Court on the ground that the plaintiffs failed to allege adequate facts or substantiate their claims with sufficient evidence. In cases where wholesale copying could not be established through adequate evidence, the courts in the US have refused to find infringement. The issue merits an investigation into how web scraping can be tackled more effectively under the existing international copyright regime. Accordingly, the paper seeks to address the following questions: (a) Whether AI-generated content based on copyrighted works qualifies as derivative works? (b) Can web scraping be protected under the fair use exception? (c) In the event of infringement, can the AI owner be held liable for the activities of AI in using copyrighted information?