Google and OpenAI’s AI video generation tools have drawn attention for the outputs they produce, particularly when tasked with creating a time-travelling doctor who operates from a blue British phone booth. The similarities to the television show “Doctor Who” raise questions about originality and copyright. Both companies’ generative AI technologies, which aim to produce novel content, may inadvertently rely on existing copyrighted works, leading to concerns about intellectual property rights.
Creative professionals, including authors and artists, are raising objections to the potential misuse of their work in training AI models. They demand compensation and permission before their material can be utilized. Notably, some publishers like the Financial Times and Guardian Media Group have entered into licensing agreements with OpenAI to address these concerns.
The predominance of protected content in these models complicates the evaluation of originality. Platforms like Vermillio have begun tracking the use of copyrighted materials in AI-generated outputs. When inputting prompts about “Doctor Who,” results revealed an 82% match to a copyright fingerprint created by Vermillio, suggesting significant reliance on existing intellectual property. Similarly, outputs from OpenAI’s Sora tool matched 87%.
Generative AI models are trained on vast datasets sourced from the open web, encompassing various content including articles and videos. As the conversation around copyright in AI progresses, the stakes increase for both creators and tech companies. For example, Anthropic recently settled a lawsuit over the use of copyrighted works to train its chatbot.
Despite these issues, Google maintains that its policies respect intellectual property rights and emphasizes that its AI models rely on publicly available data, arguing that this complies with fair use. However, there are ongoing debates, particularly in the UK, about proposed copyright law changes favoring AI development. Industry leaders are calling for a careful examination of how AI companies utilize creative content, advocating for a more equitable framework.
Source: https://www.theguardian.com/technology/2025/oct/18/the-platform-exposing-exactly-how-much-copyrighted-art-is-used-by-ai-tools