InfoSecBulletin Cybersecurity for mankind
The recent finding shows how powerful Mythos is: the AI can access the US government’s secret networks in just a few hours.
The Mythos model created by Anthropic was able to breach “almost all” of the US National Security Agency’s classified systems in a matter of hours rather than weeks, according to Senator Mark Warner (D-VA), vice chairperson of the Senate Intelligence Committee, as reported by the Financial Times.
Citrix NetScaler SAML 0-Day Flaw Under Attack
Major Danish university breached, 200,000 users at risk
Microsoft’s X account hijacked to promote Clippy crypto scam
Critical cPanel, GitLab AI Gateway and Dell CSM Flaws Enable RCE And Admin Hijacking
Nearly 100,000 email addresses exposed in first AI-related data breach in Singapore
Hackers Exploit Zimbra Mail Servers: TeamViewer patched 5 critical flaws
Google Warns of Hackers Actively Exploiting Citrix 0-Day Flaws
CISA Warns Critical MikroTik RouterOS Flaw While Cisco SD-WAN Zero-Day Exploited in Attacks
Apple Zero-Day Exploited: Pentagon Data Breach Reportedly Exposes Sensitive Data of 3 Million People
JadePuffer Agentic AI targets and destroys Azure’s cloud resources
During the briefing on June 11, General Joshua Rudd, the head of the NSA and US Cyber Command, said something that The Economist published later. We still don’t know the full details of the compromise. The specific systems affected and the methods Mythos used are not known.
Many organizations worldwide rely on Anthropic’s Mythos models, which are promoted as very strong AI tools for security. Cyber experts have said that the hacking attempt was a planned test aimed at finding weaknesses, not an illegal attack.
A well-known example was a quick use of a 17-year-old bug in FreeBSD that let code run from a distance after just one command.
Since the deal between Anthropic and the Pentagon went wrong, the US has been very active. One new change is the ban on selling Mythos-class models to other countries. Anthropic chose to stop Claude Fable 5, a restricted Mythos, because it could not check a person’s nationality before using its advanced models.
Anthropic strongly disagreed with a long-term recall of its Mythos-class models. The company argues that stopping its advanced features over minor compliance issues will slow down AI development in the West. This could make important systems easier for foreign rivals to attack, as they face fewer legal problems.
Before the Mythos-class models, Anthropic is working with federal regulators and the White House to create a risk-management plan.
Global governments are trying to stay out of geopolitics while getting advanced technology for themselves. They are pushing for more focus on local AI and tech, asking local companies to put all their effort into building their own AI systems.
Related news:
Anthropic’s Claude Mythos Preview Reveals 10,000+ 0-Days
Anthropic confirms Claude Mythos-class models will be public
