NewsBite analysis · AI, Security
The Economic Times
US cyber agency is using Anthropic's Mythos to audit government code, sources say
The US Cybersecurity and Infrastructure Security Agency (CISA) has begun using Anthropic's AI model Mythos to perform audits on government code repositories, aiming to locate potential vulnerabilities that could be exploited by cybercriminals or foreign spies. Despite ongoing tensions with the White House, government agencies appear eager to harness AI technologies to enhance cybersecurity. Sources indicate that the audits have already uncovered numerous vulnerabilities, although specific details were not disclosed.
MonitorConfidence: MediumShort term horizon
Published 7 Jul 2026, 20:30 UTC3 min read
Visit original article ↗Lead image — The Economic Times
Summary
TL;DR
- CISA employs Mythos for auditing government code.
- Focus on identifying vulnerabilities to prevent cyber threats.
- The initiative reflects growing interest in AI tools by government.
- Anthropic has faced regulatory friction with the US government.
- Mythos has reportedly shown effectiveness in detecting security flaws.
Interesting facts
- The relationship between Anthropic and US government agencies has been complex and fraught with tension.
- Mythos was previously subjected to a supply-chain risk designation due to concerns over its cybersecurity applications.