Tag
This paper proposes the Mecellem semantic protocol, an ontologically grounded framework for artificial legal intelligence, arguing that legal reasoning requires dynamic, context-dependent meaning construction rather than mere codification or statistical pattern recognition.
Legora introduces the Legora BAR benchmark for evaluating AI agentic reasoning in legal workflows, using real cases and the Legora harness to measure system-level performance.
Introduces LegalCiteTrust, a benchmark for evaluating citation trustworthiness in Chinese long-form legal research reports, assessing coverage, support, and citation fidelity. Experiments show that retrieval tools improve evidence support without reliably improving trust scores, highlighting the need for citation-aware governance after retrieval.
This paper introduces AILQA, an AI-driven system for legal question answering in the Indian legal context, using embedding and generative models with retrieval-augmented generation (RAG). It evaluates performance on Indian legal texts and the All India Bar Exam, finding that AI-generated responses can sometimes outperform reference answers.
This paper presents adaptive pipelines for legal retrieval, entailment, and judgment prediction tasks in the COLIEE 2026 competition, using multi-stage retrieval, reranking, and LLM-based reasoning.
LangSmith highlights Finch Legal, a startup using AI agents for pre-litigation in personal injury law, achieving 10x growth and utilizing LangSmith for production observability.
Ivo Benchmarks is a legal AI tool that goes beyond simple PDF summarization by pulling an entire company's negotiation history into contract review. It scores each clause based on past handling, enabling more informed decisions during contract negotiations.
This paper investigates multi-agent deliberation methods for legal reasoning tasks using LLMs, introducing two novel frameworks inspired by courtroom procedures. The experiments show that multi-agent systems achieve comparable overall performance to monolithic LLMs but produce distinct answers and can solve cases that baselines fail, highlighting the potential of multi-agent approaches for legal AI.
Awesome Legal Skills is an open-source collection of over 139 AI agent instructions and workflows for the legal industry, covering privacy compliance, contract review, litigation analysis, and more. It is compatible with mainstream AI tools like Claude Code and Codex, helping legal professionals turn their expertise into reusable AI workflows.
Harvey is experimenting with inference-time model routing and blended intelligence techniques to improve agent performance in legal tasks, highlighting both promise and risks.
Legora, an AI startup from Y Combinator's W24 batch, is defining legal AI with the launch of Cooley GO Lab using its portal, helping founders navigate legal needs.
This paper presents an architecture that uses formally verified law as a reward signal for training legal AI, adaptively autoformalizing legal rules into a formal calculus and employing a verifier to ensure provable correctness, demonstrated on German and US law examples.
Harvey partnered with Applied Compute to train a legal agent, optimizing the agent stack and post-training the GLM-5.1 model using reward signals from their Legal Agent Benchmark.
LangSmith Spotlight highlights patlytics, an enterprise legal AI platform for the patent lifecycle that uses LangSmith to orchestrate its AI stack from prompt management to workflow evaluations.
Philipp Comans shared at the Interrupt conference how Chime balances product velocity with compliance by having legal and compliance teams co-write evaluation systems, transforming AI assistant development from an 'oops-driven' approach to a continuous alignment flywheel.
This paper proposes a human-on-the-loop orchestration framework for AI-assisted legal discovery, introducing a taxonomy of agentic failures and a four-layer verification architecture to reduce privilege-waiver risk.
LegalWorld is a life-cycle interactive environment that models Chinese civil litigation as a causally connected state chain across five stages, paired with LongJud-Bench for evaluating legal agents across the full process.
Harvey CEO Winston Weinberg proposes that in the AI era, companies no longer sell software, but intelligence. Using Harvey legal AI as an example, he explains that what it truly sells is a lawyer's judgment and research capabilities.
This paper introduces LegalHalluLens, a framework for auditing hallucinations in legal AI, providing typed hallucination profiles and a Risk Direction Index to improve trustworthy deployment.
Introduces LOCUS, a comprehensive corpus of U.S. local ordinance codes designed to enable machine-readable legal AI research, covering codes from 9,239 cities and counties with ModernBERT-based classifiers for analysis.