U.S. District Court, Southern District of New York
In the highest-profile AI copyright case, Judge Sidney Stein affirmed a magistrate judge's order compelling OpenAI to produce a full 20-million-record sample of ChatGPT conversation logs to plaintiffs, rejecting OpenAI's attempt to limit production to only conversations implicating specific copyrighted works. This ruling follows the court's earlier March 2025 decision allowing the Times' core copyright infringement claims to proceed to trial. The discovery order has major implications for AI companies' data retention and litigation exposure, as it establishes that courts may require broad production of user interaction data in copyright disputes.
U.S. District Court, Northern District of California
In this landmark AI copyright case, Judge William Alsup ruled on summary judgment that using books to train AI was fair use if legally acquired, but denied fair use for works obtained through piracy. After certifying a class of approximately 500,000 copyrighted works with potential statutory damages exceeding $70 billion, Anthropic agreed to a historic $1.5 billion settlement — the largest in U.S. copyright litigation history. The settlement requires Anthropic to destroy pirated training data and pay $3,000 per work. Final approval hearing is scheduled for April 2026. This case establishes a critical precedent distinguishing between lawful and pirated training data in AI development.
Alibaba's Wan3.0 AI video generation model has entered general availability with capabilities to generate 30-second video clips from multiple input types (text, images, video, audio, web pages, PDFs, and slide decks). The model's promotional materials feature a video of Tom Cruise, raising copyright and liability concerns similar to previous AI-generated deepfake disputes involving the actor's likeness.
A copyright scholar discusses the ownership and protection of AI-generated content under European intellectual property law, examining questions about human authorship requirements and the legal status of machine-generated works under copyright frameworks.
This article addresses AI risk management and safety requirements, specifically proposing an insurance-based liability framework for frontier AI safety testing. It discusses how cybersecurity testing of advanced AI models creates third-party risks and suggests a regulatory approach modeled on nuclear power insurance liability structures to balance innovation with accountability. The piece engages with governance and compliance considerations for managing AI development risks.
This article discusses copyright litigation involving AI training on music data and AI-generated content ownership, with specific reference to a lawsuit filed in California federal court. While it opens with a question about Indian copyright law readiness, the substantive focus is on US litigation and global copyright disputes involving AI music companies—core compliance and legal topics.
This article addresses AI regulatory frameworks and international governance approaches, comparing US deregulatory policy shifts with China's regulatory model. It directly engages with AI governance strategy and international policy divergence.
AI's rapid advancement in biotechnology—including AI-assisted design of novel viruses—is outpacing existing regulatory frameworks for both AI and biological research. The development highlights gaps in governance structures needed to manage dual-use risks and biosafety concerns, prompting calls for updated regulatory oversight of AI applications in high-risk biotechnology domains.
The House Legislative Counsel's office, which drafts US legislation, is experiencing increased use of AI tools like Claude and ChatGPT by congressional staffers to generate bill language. According to Congressional lawyers, reviewing and correcting AI-drafted bills consumes more time and resources than traditional drafting processes, raising questions about the efficiency and quality control of AI-assisted legislative drafting.
This article describes a copyright agreement between ByteDance and the Motion Picture Association (MPA) addressing AI training data and intellectual property protections for generative AI models. This falls within the scope of AI-related copyright disputes and compliance frameworks governing AI-generated content and training practices.
Congressional requests for AI-related legislation surged 72% in 2025 compared to the same period in 2023, indicating accelerating legislative activity around artificial intelligence governance and regulation in the U.S. The increase reflects growing legislative interest in establishing AI-related policy frameworks and oversight mechanisms.
Round Hill Music filed a lawsuit against Anthropic and Suno alleging they used 500 copyrighted songs without authorization to train their AI systems, including compositions by James Brown and The Kinks. The case addresses core questions about copyright liability, fair use defenses, and permissible uses of copyrighted works in generative AI model training.