2 min readEnglish
Claude's AI-generated text now carries a watermark, plus more platform updates
Here is a roundup of what matters this week for teams working with Claude and AI agents: a major new watermarking feature for AI-generated text, a large investment in training engineers to deploy AI safely, and a real-world example of how a company is using AI to speed up work without cutting corners on security.
Claude Platform
Claude now adds an invisible watermark to AI-generated text
Text produced by Claude Opus 4.6, Opus 4.5, Sonnet 4.5, and Haiku 4.5 now carries Anthropic's watermark on supported platforms, a hidden signal that helps identify AI-written content. This matters if your team needs to track or disclose AI-generated material, for example in client communications or compliance reviews. Note that the watermark may not be present if certain advanced settings (temperature, top-p, top-k) are changed from their defaults, which is typically a developer-level setting.
Anthropic
Anthropic commits $100M to train 10,000 engineers for safe AI deployment
Anthropic launched Claude Frontier Academy, a $100 million program to train 10,000 Frontier Deployed Engineers by the end of 2027, aiming to bring them up to the same standard as Anthropic's own engineers. For managers, this signals a growing push toward formalized expertise in deploying AI responsibly inside organizations, which could shape future hiring and training expectations.
OpenAI
Revolut cuts coding time by 30-60% using OpenAI models and Codex
Revolut shared how it uses OpenAI models and the Codex tool to cut coding time by 30 to 60 percent, scale up financial crime investigations, and innovate while keeping security standards high. This is a concrete example of how AI can speed up technical work without compromising oversight, useful context for managers evaluating similar tools for their own teams.
OpenAI offers free access to frontier AI models for 200,000 researchers
OpenAI is expanding its ChatGPT for Academic Researchers program to give 200,000 researchers worldwide free access to its frontier models. This broadens who can use advanced AI tools for scientific work, which may be relevant if your organization partners with academic or research institutions.
OpenAI partners with NATO and allied agencies to strengthen cyber defense
OpenAI is working with NATO and allied cyber defense agencies, helping trusted teams find vulnerabilities and test fixes to protect essential systems. This shows AI being applied to national-level security work, which may be of interest for organizations thinking about how AI fits into their own security strategy.
OpenAI expands tools and partnerships to fight online child exploitation
OpenAI introduced new tools, guidance, and partnerships aimed at preventing, detecting, and responding to online child sexual exploitation and abuse. This reflects ongoing efforts by AI providers to build in safety safeguards, relevant context for organizations concerned with responsible AI use.
Summary prepared by Wisly AI from the official sources linked above. If something affects you, open the source before deciding.