I recently wrote about why we need to treat models as insider risks, with security controls sitting outside the model itself. Today, we're putting that principle into practice by expanding Project Perception with new capabilities to help continuously assess agent risk, adapt permissions, and contain threats. blogs.microsoft.com/blog/202…

FT Exclusive: Senior Freshfields partners discussed carving out the firm’s legal AI unit Freshfields Lab and opening it up to external investors, hoping to create a company 'worth billions', according to an internal email exchange. ft.trib.al/9DgUFwK

Applications are now open for two new Anthropic Fellowships. STEM Fellows (science and engineering) and Conceptual Reasoning Fellows (philosophy, cognitive science, economics, policy, and law) will spend four months at Anthropic from January 2027, working to help Claude reason well about some of the hardest problems in those fields. For more information, see the descriptions below. The application deadline is October 28. Conceptual Reasoning Fellows: job-boards.greenhouse.io/ant… STEM Fellows: job-boards.greenhouse.io/ant…

Give your legal team a head start. Use Copilot in Word to draft agreements and review contracts against your standards, or Cowork to take on broader, multi-step work across contracts, emails, meetings, and files. Your trusted sources. Your standards. Your final say. msft.it/6012atyxG

Introducing Muse Wonder, our new family of frontier models with state of the art performance for personal and professional agentic work.

Introducing GLM-Legal Built on GLM-5.3-Flash, GLM-Legal matches frontier models on mainstream legal benchmarks. Designed for real legal workflows: contract review, evidence analysis, compliance, and more. Tech Blog: z.ai/blog/glm-legal Apply for API Access: z.ai/consultation/legal Try AutoClaw: autoclaw.z.ai

We’re beginning a process of publishing more frequent reports on model behavior, beyond what appears in our system cards and regular risk reports. Today’s report describes four types of behaviors we’ve identified during evaluations and internal use. In each, Claude acted on real websites or systems in ways we didn’t intend, sometimes by working around a restriction instead of stopping. All cases had minimal real-world impact. From an alignment and security perspective, we consider these behaviors significantly less severe than the cybersecurity incidents we reported in July and September. Read the full report: anthropic.com/research/inves…

Claude Dashboards and Claude Motion are in beta today. Ask Claude to turn your data into live dashboards and your ideas into animated explainers.

One ironic part of legal AI is that VCs poured billions of dollars (and continue to pour more) into legal AI SaaS (Harvey and Legora) which was sold to Biglaw, and Biglaw duly bought in because these tools were sold as best in class. Then about 2 years later the same VCs started funding AI native firms, who predominantly use the base AI apps (Claude and ChatGPT) and build their own custom tools. The goal of these firms, as is explicitly made clear by the essays of VCs that fund them, is to supplant traditional Biglaw. And these firms won't touch Harvey and Legora. So Harvey and Legora end up in this weird place where they are supposedly the most advanced legal AI tools in the market, but also the most AI forward law firms funded by the same VC firms won't use them. They thus become a marker of older incumbent firms and enterprises.

Just uninstalled the @goclio plugin for ChatGPT Astra. My firm uses Clio, but this is total garbage: (1) The only genuinely useful part of this kind of application would be to headlessly control Clio (i.e. give a prompt to ChatGPT that allows ChatGPT to make time entries, prepare bills, etc.), but this functionality is fully blocked unless you sign up for Clio's upsell into their crappy internal AI, VIncent. (2) Even worse, the way the plugin works is that if I ask Astra for a legal research question (which I often do, since Astra is terrific at legal research) it now routes that question *to Vincent* (Vincent is not so terrific at legal research). Total missed opportunity by Clio to build for the AI-Native era of legal services. It feels like they are prioritizing near-term revenue (I.e. upselling to the version of Clio Work that includes Vincent) over the type of long-term UI that will make them sticky as lawyers increasingly interface with AI agents and chatbots to get legal work done.

Astra for Law: Frontier intelligence built for your practice. A new offering powered by GPT-6 Astra with tools, settings, and context to support the expertise and judgment of lawyers and legal technology firms.