Generative Artificial Intelligence (AI) models such as ChatGPT, DALL·E, Midjourney, Sora and Stable Diffusion has dramatically reshaped how creative content is produced. These systems can generate text, images, music, and code at scale, often indistinguishable from human‑created works. While this technological leap offers enormous benefits, it has also triggered significant intellectual property (IP) concerns, prompting legal debate and litigation globally.a
The first concerns, how we think about the training of such models whether the creators or owners of the data that are “scraped” (lawfully or unlawfully, with or without permission) should be compensated for that use. The second question revolves around the ownership of the output generated by AI, which is continually improving in quality and scale. These topics fall in the realm of intellectual property, a legal framework designed to incentivize and reward only human creativity and innovation
Most copyright regimes, including those influenced by international frameworks such as the Berne Convention for the Protection of Literary and Artistic Works, assume that a human creator is responsible for producing an original work. Generative AI complicates this principle because the output is produced by an algorithm rather than a person.
Courts have increasingly taken the view that works produced without human involvement may not qualify for copyright protection. In Andersen v. Stability AI Ltd, a group of artists sued Stability AI, Midjourney, and Deviant Art for allegedly using billions of copyrighted images without permission to train the Stable Diffusion AI model. The artists argued that the AI system unlawfully copied their works and replicated their artistic styles. The court dismissed some claims but allowed parts of the copyright infringement case to proceed, making it an important case on whether training generative AI with copyrighted material violates copyright law.
Moreover, the Kenya Copyright Tribunal in Aryeh Movement Limited v. Cynthia Beldina Akoth Okello (2025), held that works generated solely by AI cannot be copyrighted under the Copyright Act (Cap. 130), as copyright protection requires human authorship and originality. The Tribunal emphasized that only when a human contributes sufficient creative input to an AI-assisted work can it qualify for protection.
In a notable clash, Getty Images sued Stable Diffusion for allegedly using its photos without permission, claiming both copyright and trademark infringement. In November 2025, the High Court dismissed the copyright claims, ruling that the AI did not store or reproduce Getty’s images and therefore did not infringe. However, the court did find limited trademark infringement where AI-generated images replicated distorted versions of Getty’s watermark.
Furthermore, Google previously defended its use of book text for search indexing under the principle of transformative use, a precedent that continues to influence such cases.
Mitigating the Risk
This new paradigm means that companies need to take new steps to protect themselves for both the short and long term. AI developers, for one, should ensure that they follow the law regarding their acquisition of data being used to train their models. This should involve licensing and compensating those individuals who own the IP that developers seek to add to their training data, whether by licensing it or sharing in revenue generated by the AI tool (e.g., Shutterstock’s Contributor Fund), adopting revenue-sharing models (like YouTube’s Content ID), and allowing content owners to opt out of data scraping, thereby recognizing the intellectual input behind training datasets.
Organizations should conduct due diligence on the origin of their training datasets, maintain clear records of how data was obtained, and establish legal review procedures before deploying AI systems. These internal safeguards can help reduce the risk of intellectual property disputes.
By adopting these measures, companies can continue to innovate with AI technologies while ensuring that the rights of creators are respected and protected.