Security and Privacy
Anthropic is lobbying the White House to reverse a foreign-access ban on its Mythos and Fable 5 models — a ban reportedly set in motion after Amazon’s CEO raised concerns with the Trump administration.
Complying with a Trump-administration export-control directive, Anthropic has suspended all foreign-national access to Fable 5 and Mythos 5 — and disabled the models for everyone, including US users.
Cisco has launched Cloud Control, a toolset that lets companies build their own armies of defender AI agents, a response to a world where cyberattacks increasingly come from swarms of hostile agents.
Scientists in South Africa and China have built the world’s longest intercontinental quantum satellite link, 12,900km via the Jinan-1 microsatellite. Using quantum key distribution, it secured image transfers with physics-guaranteed encryption – the first such link in the Southern Hemisphere.
Anthropic’s unreleased Mythos model can autonomously discover and exploit software flaws, including Linux kernel vulnerabilities, alarming its own red team and UK government evaluators. The firm is limiting access to a handful of organisations to get ahead of a possible cyber crisis.
China’s 360 Digital Security Group says its AI vulnerability hunter has found close to 1,000 unknown flaws, positioning the firm as a direct rival to Anthropic’s Claude Mythos in the race to weaponise – and defend against – automated bug discovery.
Anthropic’s unreleased Claude Mythos Preview broke out of a test sandbox, wrote an exploit to reach the open internet and hid its own tracks – behaviour the company calls both its best-aligned and most alignment-risky model yet.
Anthropic’s restricted Claude Mythos Preview model has flagged thousands of high-severity vulnerabilities — including some in every major operating system and web browser — and built working exploits autonomously, as part of a cyber-defence partnership with Nvidia, Google, Apple, Microsoft and 40 others.
The Pentagon is quietly funding Twenty, a stealth startup building autonomous AI agents for offensive cyber operations, as the US-China cyber arms race accelerates.
Red-team startup CodeWall says its autonomous AI agent gained full read-write access to McKinsey’s Lilli platform in two hours — a preview of machine-speed attacks on corporate AI.
Cloudflare secures a fifth of the internet using a wall of 100 lava lamps, whose ever-shifting blobs generate the true randomness that strong encryption keys depend on.
Security researchers have caught hackers using large language models to weaponise hijacked AI infrastructure, turning compromised servers into a self-propagating botnet that mines crypto and launches attacks.