Claude Now Watermarks AI-Generated Content, Joining Google and OpenAI

TLDR: Anthropic now embeds an invisible watermark in text generated by new Claude models and signs image files using the C2PA standard, a move tied to the EU AI Act's transparency rules. Google, OpenAI, Meta, Microsoft and others have made similar commitments, so this is quickly becoming standard practice across the industry rather than something unique to Claude.

Why this matters: As AI writing and images spread across the web, schools, employers and platforms increasingly need a way to tell what came from a model versus a person. Watermarking gives that a technical foundation, but it is not foolproof. Knowing how it works, and where it falls short, helps you understand what these labels can and cannot prove.

Claude starts marking its own text

Anthropic confirmed this week that Claude models launched on or after August 2, 2026 automatically embed a machine readable watermark into the text they generate. The signal is invisible to readers, does not change the meaning or quality of the writing, and is designed to survive copy and paste. The change applies across Claude.ai, the API, Claude Code, Claude Cowork and Claude Tag, and Anthropic says it applies worldwide rather than only to users in the EU.

For images and other supported files, Anthropic is using a different approach: signed provenance metadata based on the C2PA standard, an open framework already used by Adobe, Google and Microsoft to record how a piece of content was created or edited.

Why now: the EU AI Act

The push comes from Article 50 of the EU AI Act, which took effect August 2 and requires AI providers to mark generated content so it can be identified downstream. Anthropic signed the EU's Code of Practice on AI generated content transparency, and according to reporting from Quartz, it is far from alone. Black Forest Labs, Google, Meta, Microsoft, OpenAI and Synthesia have all made similar pledges under the same code.

That last point is worth underlining for anyone assuming this is a Claude specific feature. Google has used its SynthID watermarking system in image and text tools for a while, and C2PA metadata is already common practice among major AI image generators. Anthropic's announcement effectively brings Claude's text output in line with where the rest of the industry is heading.

What the watermark does not prove

Anthropic is upfront that a detected watermark is not conclusive proof of authorship. Claude may have simply proofread or translated a human's own writing, and heavy editing, format conversions or asking a different model to rewrite the text can strip the signal entirely. The company has also not yet released a public detection tool, so for now there is no way for an outside reader to check a piece of text against the watermark directly.

Still, this is a meaningful first step toward giving platforms and regulators a consistent signal to work with, even if it is not a perfect one.

Source: TechCrunch

How to Configure Claude Desktop for Your Enterprise: A Complete Admin Guide

If you're rolling out Claude Desktop across your organization, you don't have to leave settings up to individual users. Claude Desktop now supports enterprise-grade configuration for Team and Enterprise plan admins, letting IT teams manage the app centrally through familiar tools like Microsoft Intune, Jamf Pro, and Group Policy.

Manage Claude Desktop Through Your MDM

On macOS, Claude Desktop reads its settings from the com.anthropic.claudefordesktop preference domain, which means you can deploy configuration profiles using any standard MDM tool — Jamf Pro, Kandji, Intune, or third-party profile editors like ProfileCreator and iMazing Profile Editor. This lets admins push settings to specific machines or user groups without needing anyone to touch a settings menu.

On Windows, configuration happens through Group Policy or Intune, with settings applied at either the machine level (HKLM) or user level (HKCU). If you set policies at both levels, machine-wide settings take priority, giving admins a reliable way to enforce standards across the fleet.

What You Can Control

Enterprise policies cover a practical range of controls, including:

• Auto-updates: Disable them if your MDM manages app versioning, or set how many hours (up to 72) before an update is force-applied.
• Organization login restrictions: Require users to log in under specific organization UUIDs, blocking unauthorized accounts from accessing Claude Desktop.
• Extensions and MCP servers: Enable or disable desktop extensions, the extension directory, and local MCP servers depending on your security posture.
• Cowork and Claude Code access: Turn these features on or off, and restrict which workspace folders can be mounted.

These settings are managed through registry keys on Windows (under HKLM:\SOFTWARE\Policies\Claude) or configuration profiles on macOS, giving IT teams a consistent way to enforce policy regardless of platform.

One Thing to Watch: The Allowlist Override

There's an important interaction worth knowing before you deploy: enterprise policy controls set at the machine level will override the in-app desktop extension allowlist. If your organization relies on that allowlist, make sure isDesktopExtensionEnabled and isDesktopExtensionDirectoryEnabled aren't set to "false," or the allowlist won't be able to populate its registry properly.

Why This Matters for IT Teams

Centralized configuration means fewer support tickets, more consistent security policies, and less risk of shadow IT creeping in through unmanaged extensions or MCP servers. For organizations already using MDM to manage other desktop apps, adding Claude Desktop to that workflow should feel familiar rather than like learning a new system.

If you're planning a broader rollout, it's worth pairing this configuration guide with Anthropic's platform-specific deployment articles for macOS and Windows to make sure installation and policy management are aligned from day one.

Source: Enterprise configuration for Claude Desktop – Claude Help Center

Claude Code iOS Simulator: Build and Test iOS Apps in Real Time (Public Beta)

If you build iOS apps, you know the rhythm: write code, switch to Xcode, wait for the build, tap through the simulator, spot the bug, switch back, repeat. Anthropic just closed that loop. As of today, July 21, 2026, Claude Code Desktop can open, run, and interact with your app directly inside the iOS Simulator, without you ever leaving the conversation.

The feature is a new iOS Simulator pane. When you ask Claude to build, run, or check your app, the pane opens automatically next to your chat and streams the simulator's screen live, so you can watch Claude tap through your onboarding flow, install a build, or verify a bug fix in real time. It's interactive too: you can tap and swipe on the device yourself, use standard simulator shortcuts, switch between devices, and save a screenshot or recording. Because it drives the simulator directly, it skips the macOS permissions that Claude's other screen control tools need. Each session gets its own simulator, and you can have up to four panes open at once.

The real value isn't just watching Claude work, it's that Claude can now close its own feedback loop. Instead of writing code and telling you it should work, Claude can build the app, launch it, tap through the flow it just changed, and catch its own mistakes before handing things back to you. That's a real shift from "AI that writes code" to "AI that verifies its code actually runs."

You stay in control throughout. The first time Claude touches a simulator, the app asks for one time permission. Opening a URL on the device or running a build always follows your session's permission settings, since those actions can move data or execute scripts. Screenshots Claude takes are sent to Anthropic and retained under your normal conversation settings, so it's worth avoiding real logins on a simulator Claude is driving.

This is a public beta, live today in Claude Code Desktop on macOS for Pro, Max, and Team plans. You'll need Claude Desktop v1.24012.0 or later and Xcode with the iOS platform installed, and it only works in local sessions, not cloud or SSH. It's simulator only for now, so testing on physical devices still means going through Xcode yourself.

If you're already using Claude Code for iOS work, it's worth updating for right away. If you're on Windows, Linux, or working purely in the cloud, keep an eye on whether the pane expands beyond local macOS sessions.

For full setup steps and troubleshooting, see Anthropic's docs: Test iOS apps in the simulator.

Anthropic Launches Claude for Teachers

Anthropic announced Claude for Teachers on July 14, 2026, giving verified K-12 educators in the US free access to premium Claude capabilities, a library of teaching-specific skills, and a direct connection to evidence-based curricula mapped to academic standards in all 50 states.

The idea behind it: research shows that practices like differentiated instruction and mastery-based learning improve student outcomes, but teachers rarely have the time to implement them well — especially in under-resourced schools. Claude for Teachers connects to Learning Commons for standards alignment, pulls in curricula like OpenSciEd and Illustrative Mathematics, and integrates with existing classroom tools like Canva Education, MagicSchool, and Diffit. Teachers can ask it to draft standards-aligned lesson plans or differentiate materials across readiness levels, and — since it includes Claude Code and Cowork — it can also handle longer-running tasks like analyzing class data or automatically adjusting tomorrow's lesson plan based on the day's exit tickets.

On the privacy side, it comes with its own teacher terms and a K-12 Data Processing Addendum built for FERPA compliance, and Anthropic says the data isn't used for model training. They're also working with the American Federation of Teachers on privacy and safety standards for K-12 AI tools.

Access is currently for individual educators (a district/school offering is coming later), and verified teachers can sign up through June 30, 2027 for a free year.

Full announcement: anthropic.com/news/claude-for-teachers