AI × Law

John Nay says Norm AI and Norm Law are built around “agentic law” — AI agents that do first-pass legal and regulatory work in the background, then learn from prior matters. He says Norm Law uses lawyers, legal engineers, software engineers and AI engineers, with outcome-based pricing and final review by a barred attorney.

Legal AI startup Vikk AI says it raised $4.2 million in a pre-seed and seed round to expand its consumer legal AI platform and sell sponsored placements to law firms. The company also plans a lawyer directory and an enterprise data product built from anonymized legal inquiry data, signaling growing competition around how consumers find legal help online.

A July 21, 2026 firm update surveys emerging AI legal risks across patent eligibility, trade secret theft through AI tools, privacy, data scraping, AI wiretapping, copyright, products liability, Section 230, employment, and securities. It’s not a ruling or product launch, but it gives litigators a useful snapshot of where AI disputes and enforcement are headed.

A federal judge in San Francisco granted final approval to a $1.5 billion settlement between Anthropic and a class of authors over alleged illegal downloading of copyrighted books. The court called it the largest copyright class action settlement in history; the deal resolves past data-acquisition claims, but not future claims over model outputs.

GC AI’s July 15, 2026 article catalogs the current legal AI market and highlights several concrete developments, including Spellbook’s exclusive partnership with the Canadian Bar Association, LegalFly’s planned Series B, and new legal modes from Anthropic and OpenAI. It also says in-house generative AI usage has reached 87%, underscoring how quickly legal AI has moved from experiment to workflow tool.

Y Combinator published a directory of 47 legal-tech startups in its July 2026 batch, spanning AI tools for law firms, in-house teams, litigation support, compliance, immigration, and contract work. The list is a useful snapshot of where startup activity is clustering across the legal market, including companies building products for document review, discovery, trial prep, patent enforcement, and legal operations.

Legal tech startups kept drawing serious capital, including Courtroom’s pre-seed launch, Fearn’s $5.5 million seed, Sandstone’s $30 million Series A, Wordsmith’s $70 million Series B, Legora’s $50 million Series D extension, and Patlytics’ $40 million Series B. The rounds show continued investor appetite for AI tools spanning litigation, patent work, and in-house legal operations.

The article says Otter.ai is being sued in federal court in California over allegations it captured audio, transcripts and voiceprints from Zoom, Teams and Google Meet calls without consent and used that data to train models. It also says Fireflies.ai faces separate biometric privacy suits in Illinois, while Google Meet and Microsoft Teams are increasingly blocking or flagging third-party notetaker bots by default. That makes the piece relevant to litigators and legal ops because it ties live privacy litigation to concrete product and platform changes affecting meeting transcription workflows.

A National Center for State Courts article says a California judge detected deepfake audio and video submitted as authentic testimony in Mendones v. Cushman & Wakefield, highlighting how easy it is now to fabricate evidence. The piece says courts are responding with proposed evidence rules, California Judicial Council guidance, and NCSC bench cards for evaluating AI-generated materials.

NIST’s Center for AI Standards and Innovation issued a request for information on how to measure and improve the secure development and deployment of AI agent systems. The agency says the feedback may inform technical guidelines and best practices for AI security, including risks like hijacking and backdoor attacks.

A federal judge in Illinois rejected Apple’s bid to move eight related biometric privacy lawsuits over AI voice training to one judge. The cases accuse Apple, Nvidia, Microsoft, Amazon, and Meta of using people’s voices to train AI without permission, making it a significant early AI privacy litigation battle.

Consilio said it is a cornerstone investor in The LegalTech Fund II and is expanding its Innovation Lab to include a new offering for law firms. The move gives legal teams and firms earlier access to emerging AI, automation, and analytics tools, which is relevant for legal departments and litigators evaluating new tech for workflows and e-discovery.

The LegalTech Fund said it has closed a second venture fund at $110 million to back early-stage legal technology startups. That is a concrete financing signal for the legal-AI and legal-tech market, and it gives the firm more capital to support new companies in the space.

Legal tech saw more than $1 billion in funding announcements between mid-September and mid-October 2025. The raises spanned Filevine, EvenUp, Eve, Airia, Harvey, Spellbook, MarqVision, Enter, Nexl, Crosby, Lexroom, Lucio, PostSig, and Saga, underscoring continued investor appetite for legal AI and workflow tools.

This roundup profiles 10 U.S. legal tech startups using AI across litigation, contracts, claims, client intake, and case analysis. It also details recent funding for several companies, including Harvey, Supio, Paxton, Theo AI, Justpoint, and Caseflood.ai, making it useful for tracking where legal AI money and product development are heading.

Thomson Reuters has launched CoCounsel Legal, adding Deep Research and guided workflows built on Westlaw and Practical Law content. The company says the product is designed for litigation, transactional work, and regulatory analysis, and includes workflows for drafting complaints, discovery requests and responses, and deposition transcript review.

A Berkeley Technology Law Journal blog post argues that generative AI is creating new problems for authenticating trial evidence, from fake audio/video to baseless “deepfake” defenses. It points to inconsistent judicial responses in cases including Huang v. Tesla, Valenti v. Dfinity, USA v. Khalilian, Wisconsin v. Rittenhouse, and US v. Reffitt, and says the Federal Rules of Evidence Advisory Committee is considering changes to Rule 901.

Thomson Reuters Institute says AI-generated evidence is forcing judges to rethink authenticity and reliability, especially as deepfakes and manipulated files become harder to spot. The piece highlights new bench cards from the AI Policy Consortium and discusses recent cases, including a Washington ruling excluding AI-enhanced video and a California court rejecting a deepfake challenge to video evidence.

Legaltech Hub says Walton is a 2025 US-based transaction-management platform aimed at law firms. It coordinates deal workflows end to end, generates documents from structured deal terms, reviews diligence materials with AI, and is being built with firms for areas like M&A, fund formation, banking, finance, and real estate.

Legora says it has launched a new aOS for legal work, plus new products including Agent, Monitors, and Lists. The company says the system is built for law firms and in-house teams and highlights claimed efficiency gains such as reduced non-billable hours and added billing capacity.