AI × Law

The Oregon Supreme Court has issued its first orders sanctioning litigants for filing documents containing fabricated quotes and citations generated by AI. The court dismissed a petition for writ of mandamus after repeated hallucinations in one case and imposed a $500 monetary sanction in another.

The Eighth Circuit's ruling in United States v. Belt highlights growing judicial tensions over authenticating video evidence as AI-generated deepfakes proliferate. In response, the Federal Advisory Committee on Evidence Rules is studying proposed Rule 901(c), which would require judges to determine evidence fabrication challenges before presenting media to juries.

Securities class actions related to artificial intelligence at mid-year 2026 have already surpassed the total number filed in all of 2025, according to a report by NERA economists. The surge reflects heightened scrutiny of corporate AI disclosures, forward-looking statements, and business strategies as average settlement values hit a decade high.

A federal court in California dismissed a Defend Trade Secrets Act claim against OpenAI after the plaintiff built her proprietary frameworks using ChatGPT, effectively waiving confidentiality. The decision underscores how using consumer AI tools without contractual data isolation can destroy trade secret protections and legal privilege.

ClaimHit has introduced an AI patent intelligence platform that maps patents to real-world products and standard specifications. The tool uses a six-channel discovery process and an ensemble of six LLMs to generate evidence-anchored claim charts and essentiality matrices for IP litigators and strategists.

In litigation over a lethal Houston explosion, plaintiffs' counsel successfully subpoenaed the ChatGPT transcripts of an expert witness hired by 3M. The logs revealed the expert relied exclusively on ChatGPT to write his 30-page report, prompting AI with instructions to show 3M was 0 percent at fault.

The New York City Bar Association released ethics guidance stating attorneys who have permission to record nonclients may ethically use AI to transcribe and summarize those conversations. However, the committee warned that lawyers must still ensure confidentiality and evaluate whether using AI tools is advisable given potential risks.

Harvey has introduced Harvey II, the next generation of its legal AI platform. The updated technology is designed to integrate artificial intelligence more deeply into attorneys' daily workflows.

Blackstone-backed Norm Ai and its affiliated hybrid law firm, Norm Law, have raised over $250 million to target regulatory compliance work traditionally handled by Big Law. The firm leverages agentic AI workflows and legal engineers to offer flat fees that directly undercut traditional billable hourly rates.

Legal AI startup Harvey has announced its first in-house language model, Harvey Tenet, designed to handle complex legal reasoning while reducing reliance on external model providers like OpenAI and Anthropic. The launch is part of a broader 'Harvey II' platform update, which includes new memory features that allow AI agents to retain user preferences and working context.

Legal AI platform Harvey has integrated DeepL's translation API directly into its platform to power document translation across more than 100 languages. DeepL will handle over a third of Harvey's total translation volume, allowing users to translate complex legal files while preserving context, formatting, and custom terminology.

Legal tech and services provider Elevate has acquired Lupl, an AI-native legal project management and workflow platform, for an undisclosed sum. Lupl's platform, which integrates natively with Claude, Harvey, and Microsoft Copilot, will expand Elevate's legal software offerings for law firms and corporate legal departments.

Quinn Emanuel's co-managing partners revealed the litigation powerhouse is making significant investments in AI by partnering directly with Anthropic, OpenAI, SpaceX, and others. The firm is customizing these technologies specifically for its litigators while simultaneously expanding its associate hiring.

Relativity has unveiled claiR, a conversational AI tool for RelativityOne that allows lawyers to ask plain-language questions directly against matter data without transferring files or relying on technical teams. The tool is currently being trialed by major firms including A&O Shearman, Foley & Lardner, and K&L Gates ahead of a planned general release in early 2027.

In a large cross-border antitrust investigation involving 69.6 million documents, legal teams from Redgrave and Consilio successfully combined traditional TAR with generative AI. Using Gen AI as a precision filter on 1.4 million TAR-positive documents eliminated 27.5% of false positives before production, significantly reducing downstream review burden and jurisdictional risk.

A new industry survey from ComplexDiscovery and EDRM shows that generative AI-assisted review pricing is fragmenting, with hybrid and per-document models each capturing ~28% of reported approaches. Per-document GenAI review rates between $0.11 and $0.50 are emerging as a competitive zone challenging traditional human review economics.

Corporate legal departments are moving past initial generative AI pilots to establish formal governance over autonomous AI systems used for contract review, regulatory monitoring, and first-pass discovery. General counsel are increasingly prioritizing AI vendor contract addenda, liability allocation, and regulatory compliance as agentic tools take on more complex workflows.

IP firm ZeusIP, in partnership with Native AI, has launched Real-time Intelligence Agent (RIA), an AI platform designed to automate research and evidence review for trademarks, patents, and litigation. Powered by Small Language Models trained on eight million documents, the tool claims to reduce task times by up to 4.6x while keeping client data within a controlled ecosystem.

Davis Wright Tremaine LLP has announced a firmwide rollout of the legal AI platform Harvey alongside Microsoft Copilot and Microsoft's AI Frontier Suite. The firm aims to achieve a 90% adoption rate among its more than 600 attorneys and staff to streamline research, drafting, and document analysis workflows.

Swedish legal AI platform Legora is reportedly in talks to raise new capital at a valuation exceeding $10 billion, up from $5.6 billion in April. The company reports annual recurring revenue of $150 million and counts major law firms including White & Case and Cleary Gottlieb among its clients.