Two recent rulings from the Northern District of California provide fresh insight into how U.S. courts are confronting copyright challenges in the context of artificial intelligence. As developers deploy increasingly sophisticated generative AI tools, courts are now tasked with assessing how traditional copyright doctrines apply to machine-generated outputs.

In Andersen v. Stability AI, plaintiffs alleged that image-generation models like Stable Diffusion infringed on their copyrighted works by training on protected datasets. The court declined to dismiss the case in full, ruling that it raised plausible claims that AI-generated outputs may constitute derivative works. However, the decision also limited the scope of the complaint, rejecting blanket infringement assertions and emphasizing the need for specificity in demonstrating substantial similarity between outputs and protected originals.

Similarly, in Doe v. GitHub, developers sued Microsoft, GitHub, and OpenAI over allegations that their Copilot tool improperly repurposed code from public GitHub repositories without adequate attribution or compliance with open-source licensing terms. While the court dismissed some claims—including those grounded in DMCA violations—it allowed others to proceed, particularly around the question of whether Copilot’s outputs could improperly reproduce or adapt protected source code.

Both rulings point to a growing judicial willingness to scrutinize how generative AI systems operate under existing copyright frameworks. Notably, neither court accepted the defendants’ arguments for categorical exemptions based on fair use or data transformation. Instead, the judges stressed that AI-generated material must be analyzed using traditional tests for infringement, including the degree of transformation and the nature of the input data.

The broader context is one of escalating global litigation over AI’s reliance on copyrighted material. The BBC has reported on similar claims in Europe and the UK, where authors, musicians, and software developers are raising parallel concerns. The U.S. Copyright Office, meanwhile, is actively reviewing policy guidance on AI-generated works, with implications for future litigation strategies and compliance regimes.