AI × Law

Pulitzer winner John Carreyrou and five other authors rejected the $1.5 billion Anthropic class settlement to file individual copyright suits against major AI developers. A federal judge recently severed the claims into separate actions against Anthropic, OpenAI, Meta, Google, xAI, and Perplexity.

A analysis of recent AI copyright rulings, including Bartz v. Anthropic and Thomson Reuters v. Ross Intelligence, highlights how courts are distinguishing between transformative model training and the unlawful retention of pirated datasets. The cases demonstrate that data acquisition methods and direct market competition remain major vulnerabilities for AI developers.

The Arizona Court of Appeals sanctioned an unrepresented appellant in Estate of Acciavatti for submitting an AI-generated brief that contained fake case citations and misstatements of law. The court ordered the litigant to pay a portion of his sister's legal fees, reiterating that litigants are personally responsible for verifying AI-generated legal work.

Legal tech specialist Legora is seeking a $10 billion valuation amid a broader surge in venture capital funding for legal technology and AI startups.

A Massachusetts federal judge denied without prejudice a motion by major record labels to amend their copyright infringement complaint against AI music generator Suno to add 61,026 additional sound recordings. The court ruled that adding the massive volume of works near the end of fact discovery would cause unnecessary complexity and prejudice Suno's right to a timely ruling on its fair use defense.

The Ninth Circuit vacated a preliminary injunction against Perplexity AI in Amazon v. Perplexity, holding that an AI agent is a tool and the human user is the party 'accessing' a website under the CFAA and CDAFA. The decision provides the first major federal appellate guidance on anti-hacking statutory liability for agentic AI tools.

Florida's Eleventh Judicial Circuit issued Administrative Order 26-04, requiring all attorneys and pro se litigants to explicitly disclose GenAI use on court filings. Anyone using GenAI must include a specific certification confirming they independently verified all factual assertions, legal authority, and citations.

Nine tech companies—including Apple, Google, Meta, and Microsoft—face proposed class actions in Chicago federal court alleging they illegally extracted voiceprints from podcasts and audiobooks to train AI models without consent. Filed under Illinois’ Biometric Information Privacy Act (BIPA), the lawsuits mark a major shift from copyright to privacy-based litigation over AI training data.

A court ruled that an author's allegations regarding Adobe's alleged use of pirated books to train NVIDIA's Nemotron AI models were too conclusory to establish standing. The decision highlights the strict pleading standards plaintiffs face in AI copyright litigation.

An appellate court has issued what experts call the first ruling addressing agentic AI, offering a blueprint on consent and server access for AI developers. Legal experts note that while the decision is limited in scope, it signals how courts will evaluate AI agent liability and user consent going forward.

A Connecticut Superior Court judge sanctioned a self-represented litigant who embedded invisible white-on-white text in filings instructing any AI reviewing the document to rule in his favor. Judge Walter Spader Jr. revoked the litigant's electronic filing privileges, requiring future submissions on paper to eliminate hidden digital commands.

U.S. District Judge Gloria Navarro dismissed a lawsuit against a Nevada family court judge who was accused of delegating her decision-making authority entirely to AI. The court held that issuing rulings is an official judicial function protected by absolute judicial immunity, regardless of whether AI was actually used.

McDermott Will & Schulte is recruiting a dedicated Director of Attorney AI Learning & Strategy to lead the firm's AI curricula, credentialing, and practice-based training. The role will oversee cross-functional AI implementation across practice groups while addressing risk, supervision, and core legal skills retention.

A group of journalists, podcasters, and audiobook narrators filed class action lawsuits against nine tech giants including Apple, Meta, and Microsoft in Chicago federal court. Rather than relying on copyright law, the claims allege violations of Illinois' Biometric Information Privacy Act (BIPA) for using unauthorized voice data to train AI models.

Legal tech startup Harvey has launched Harvey Tenet, its first model built on an open-weight base rather than closed proprietary systems from OpenAI or Anthropic. The San Francisco firm post-trained Chinese lab Moonshot AI's Kimi K3 base model to achieve specialized legal performance at lower inference costs.

The Third Circuit has heard oral arguments in Thomson Reuters v. Ross Intelligence, setting up the first federal appellate ruling on fair use and AI training. The decision could clarify how market harm and transformative use apply to legal AI models trained on proprietary headnotes.

Indian courts are rapidly expanding the scope of personality rights to combat AI-driven threats like deepfakes, synthetic voices, and unauthorized commercial likenesses. The judiciary's evolving legal framework aims to protect public figures and private individuals from unauthorized AI exploitation in the absence of comprehensive statutory regulation.

Nine major tech companies—including Apple, Meta, Microsoft, and ElevenLabs—face potential class-action lawsuits in Chicago federal court alleging they used human voice recordings without consent to train AI models. Filed by civil rights firm Loevy & Loevy on behalf of journalists and podcasters, the suits claim violations of the Illinois Biometric Information Privacy Act (BIPA) and seek hundreds of millions in damages.

IP intelligence platform Patlytics has introduced Agent, an AI reasoning layer designed for patent litigation, prosecution, and portfolio management. Powered by 150+ pre-built skills, the system integrates with a database of 145 million patents and 40 million case law records to automate tasks like claim construction support and initial case assessments.

Clio has completed its $1 billion acquisition of legal research platform vLex while securing $500 million in Series G funding at a $5 billion valuation. The massive capital injection and acquisition significantly expand Clio's capabilities, positioning it as a major competitive force in legal operations and AI-driven research.