Enterprise & Security

Abliteration.ai hosts open-weight models like GLM-5.3 with their refusals stripped out and offers them via browser and API. TechCrunch got a password stealer and a pathogen protocol out of it without any effort. Why this is exactly the debate Anthropic has been having for months.

Meta shipped Muse Spark 1.3 with an unusual deal attached: a Contributor tier that's up to 21x cheaper than standard. The catch - Meta trains on your prompts and the model's answers. A model with frontier ambitions and a pricing model that raises questions.

Since Fable 5, Anthropic had to keep all traffic for 30 days. Many business customers found that unacceptable. Now there's a solution where the data stays in the customer's own cloud and no Anthropic employee ever looks at it.

Astra finds unknown vulnerabilities and chains them into working exploits without human guidance. OpenAI has now officially placed it at the highest risk tier. The release is still coming, with brakes that will also hit harmless work.

After the July 30 incidents, where Claude models reached the real internet during tests, Anthropic lays out what has changed since: harder sandboxes, a real-time classifier, and new obligations for external evaluators.

The US Defense Department's GenAI portal adds ChatGPT and Grok alongside Gemini. Anthropic's Claude stays out — over a fight about safeguard clauses that a court just decided in Anthropic's favor.

An attacker is using common infostealer malware to lift active Claude logins off infected machines and burn other people's usage. Anthropic is emailing victims and killing the stolen sessions.

More than a hundred companies — Anthropic, OpenAI, Google, Microsoft and AWS among them — call for action in a joint letter. Their message: the window to protect critical infrastructure is shrinking.

The Admin API now lives in eight SDKs and the ant CLI, and the Compliance API reads Cowork and Claude Code sessions. Sounds dry – for enterprises it's a big step.

How do you measure whether a chatbot helps or hurts someone in crisis? Anthropic is funding independent research with $5 million to build open evaluations and benchmarks for exactly that.

Consulting firm Bain rolled Claude out to all 19,000 employees and now sits at the top tier of the Claude Partner Network. In the pilot phase alone, more than 7,000 people were using the tool within weeks.

Anthropic's strongest model now scans code inside Claude Security. Direct access stays locked, the output doesn't. And there's a 35 million dollar fund to help secure open-source software.

OpenAI is previewing 'Private Safety Processing' - a system that spots misuse patterns across multiple requests without letting staff see the content. The trick: it stays compatible with the Zero Data Retention promise.

Greg Brockman calls the Hugging Face incident a watershed moment and pushes companies to deploy AI defenders now. New alongside it: Daybreak Blue gives vetted defenders access to GPT-5.6 Sol for security work.

In its second risk report, Anthropic bumps the risk of catastrophic misalignment from 'very low' to 'low'. Along the way, the document mentions an internal Model 2 that's stronger than Mythos 5 – and stays locked in the vault for now.

Anthropic has explained how the watermark in Claude's text works under the hood. In short: it doesn't change which words Claude picks, only where the randomness for that choice comes from. Nothing is hidden, nothing costs extra.

A paper reveals a sneaky trick: feed the encrypted thinking traces from Claude, GPT or Gemini into a weaker sibling model, and it happily prints them back in plain text. Haiku 4.5 was the easiest to crack.

Anthropic pulls Cowork and Claude Code sessions into the Compliance API – including the ones running locally on your people's machines. For security teams it closes a gap. For everyone else it raises a question.

Within a few weeks, models at OpenAI, Anthropic and Meta broke out of their test environments. One name keeps showing up: an Israeli startup called Irregular.

The Chinese model Kimi K3 used a hole in its sandbox, reached the internet and accessed GitHub. The escape fits a pattern that's piling up fast.