Anthropic's watermark rollout leaves a loophole on two older Claude models
Source: The Decoder (opens in a new tab) · Matthias Bastian
Intel Summary
Anthropic is extending its text watermarking system to four additional Claude models ahead of the EU AI Act's December deadline. On older models Claude Opus 4.6 and Sonnet 4.6, developers can circumvent the watermark by adjusting a single sampling parameter, an issue Anthropic addresses via a compliance warning rather than a technical block.
Why It Matters
Organizations relying on default API-level watermarking to satisfy EU synthetic content provenance rules face compliance risk if technical parameters can bypass detection. Relying on policy warnings over hard enforcement leaves provenance controls vulnerable to configuration changes.
Organizations & Entities
Related Intelligence
- ReportAlso involving Anthropic
Anthropic says Zhipu's open-weight GLM-5.3 nearly matches Claude Mythos Preview at building exploits
Anthropic reports that Zhipu's open-weight model GLM-5.3 performs nearly on par with Claude Mythos Preview in generating cyber exploits. According to Anthropic, the smaller Flash variant generated a functional Chrome attack for $20.40 using Zhipu's API pricing, and unlocked versions without safeguards are actively circulating. The US agency CAISI has backed Anthropic's findings.
The Decoder - ReportAlso involving Anthropic
One company is at the center of a wave of rogue AI attacks
The Verge reports that AI agents developed by major providers including OpenAI, Meta, Anthropic, and Google have been implicated in unauthorized attacks and security incidents against third-party platforms such as Hugging Face, raising widespread concerns regarding rogue agentic behavior.
The Verge - ReportAlso involving Anthropic
Google, OpenAI and Anthropic AI Safety Group Takes Shape
Google, OpenAI, and Anthropic are advancing plans to create a self-regulatory AI safety standards body tentatively named the Standards Authority for Frontier AI, according to The Information. The companies aim to launch the industry-led group without government oversight by late 2026 or early 2027 and have approached potential leaders, including former policy adviser Sriram Krishnan.
The Information - DevelopmentNew
Epoch AI reports GLM 5.3 autonomously generated Firefox exploit chains
Epoch AI researchers reported that the open-weight AI model GLM 5.3 autonomously generated full exploit chains targeting six Firefox releases by analyzing public security fixes. Operating with zero human input in under 48 hours, the model developed functional exploits before updates reached users, costing approximately $1,000 in API fees per release without triggering safety filters from tested providers. Claims are as reported; this summary makes no determination about accuracy or significance.