AI sued by image library for intellectual property infringement in training models - UK Human Rights Blog

The High Court’s ruling in Getty Images v Stability AI establishes that AI model weights do not constitute infringing copies under current UK copyright law. Although Stability AI was found liable for limited trademark infringement involving watermarks in early model versions, the core claim that training an AI on copyrighted images constitutes secondary copyright infringement was dismissed. This decision effectively separates the ephemeral exposure of data during machine learning from unlawful copying, creating a significant legal shield for developers who train models overseas. This judgment exposes critical gaps in existing intellectual property frameworks, particularly regarding the transparency and control of AI training processes. Because model weights encode learned patterns rather than storing original works, the current legal definition of an "infringing copy" fails to protect creators whose data is processed into intangible parameters. Consequently, rights holders face substantial hurdles in proving infringement, highlighting an urgent need for legislative updates to address the opaque nature of generative AI technologies. For the open data community, this case underscores the tension between unrestricted data access for innovation and the protection of proprietary rights. It signals that without specific statutory exceptions, such as an opt-in system for text and data mining, the current legal balance may favor AI developers over content creators. The ruling suggests that future open data initiatives must navigate a landscape where copyright liability can be avoided through jurisdictional strategies and technological interpretations that do not yet account for the realities of machine learning.

Source: ukhumanrightsblog.com
Published on 2025-11-08

AI sued by image library for intellectual property infringement in training models - UK Human Rights Blog

Related news