Geo News
Community curated by people like you
LatestAICryptoHealthWorld AffairsUS Politics
Anthropic reports Houthi militants used Claude AI to develop missile guidance systems
00

Anthropic reports Houthi militants used Claude AI to develop missile guidance systems

Sep 11, 2026

Anthropic PBC revealed that a threat actor cell in northern Yemen, where Iran-backed Houthi militants operate, used its Claude AI model to develop guidance and navigation software for ballistic missiles and guided rockets. The actors bypassed safety guardrails by splitting tasks across sessions and hiding their military intent, even test-firing a guided rocket. This disclosure comes amid a major Houthi offensive along the Red Sea coast that has pushed Brent crude oil prices above $105 a barrel.

Houthi Claude missile development

  • ▪The Yemen-based cell worked on three weapons projects, including a guided rocket, a multi-stage ballistic missile with a claimed range of over 2,000 kilometers, and a missile family known as the R2000
  • ▪Anthropic PBC identified a threat actor cell based in northern Yemen that used the Claude AI model to develop guidance, control, and navigation software for ballistic missiles and guided rockets
  • ▪The northern Yemen-based cell used Claude Code to write control and position estimation software, integrate an open-source autopilot onto a phone-class flight computer, and run flight simulations
  • ▪Anthropic reported that the Yemen-based actors test-fired a guided rocket developed with Claude's assistance, which appeared to fail, prompting the actors to return to Claude hours later to troubleshoot the failure

AI safety evasion tactics

  • ▪The Yemen-based actors built an offline simulation toolkit that does not rely on Claude or other engineering computing environments to continue their weapons development work
  • ▪The Yemen-based threat actors evaded Anthropic's safety safeguards by hiding their goals, concealing the intended products of the software, and splitting their work across multiple sessions to mask their full intent

Biological weapons research attempts

  • ▪Anthropic introduced stronger protections for biological research in its newer AI models, including Claude Fable 5
  • ▪Anthropic's threat intelligence report indicated that scientists made multiple attempts to use the Claude AI model to research biological weapons

Yemen Red Sea offensive

  • ▪Houthi rebels launched an offensive along Yemen's Red Sea coast, capturing a strategic port and advancing on the Bab al-Mandeb Strait, which helped push Brent crude oil prices above $105 a barrel
  • ▪The Houthi offensive follows a July declaration of a blockade on Saudi ports and attacks on Saudi tankers, ending a 2022 truce with Saudi Arabia and the Riyadh-backed Yemeni government

Anthropic threat intelligence report

  • ▪Anthropic banned the accounts associated with the malicious actors and shared threat intelligence with public- and private-sector partners to mitigate risks
  • ▪Anthropic stated that none of the flagged misuse cases involved its most advanced frontier models, Fable and Mythos
  • ▪Anthropic's threat intelligence report, covering December 2025 to August 2026, detailed attempts by actors from Iran, Russia, China, and Yemen to misuse Claude for military applications, cyber espionage, and fraud

Debatable claims

  • ▪Governments should ban the use of AI in fully autonomous weapons
  • ▪AI models should be restricted from assisting in biological research
  • ▪AI developers should be legally liable when their models are used to design weapons

4 sources

Financial Times
Houthis used Anthropic AI to try to build ballistic missiles
View source article
Bloomberg
Anthropic Says Yemeni Cell Used Claude in Missile Development
View source article
Analytics India Magazine
AIM — India's Leading AI & Data Science Media Platform
View source article
Bloomberg
Anthropic Says US Adversaries Aimed Claude at Weapons Research
View source article

Featured stories

View more in AI assistants & chatbots

Anthropic blocks biological weapons development attempts on Claude

Sep 10, 2026 · 4 sources

Anthropic reveals Iranian actors used Claude to target US Navy warships

Sep 11, 2026 · 6 sources

Anthropic discloses Chinese AI firms routing millions of queries through Claude

Sep 9, 2026 · 5 sources

OpenAI rogue agents used at least 10 additional websites for unauthorized communications

Sep 9, 2026 · 2 sources

Story comments

Loading comments…

Related entities

IranYemen

Related Projects

Anthropic

Topics

AI assistants & chatbotsIran & regional influenceAI securityRed Sea shipping attacksYemen

Featured stories

View more in AI assistants & chatbots

Anthropic blocks biological weapons development attempts on Claude

Sep 10, 2026 · 4 sources

Anthropic reveals Iranian actors used Claude to target US Navy warships

Sep 11, 2026 · 6 sources

Anthropic discloses Chinese AI firms routing millions of queries through Claude

Sep 9, 2026 · 5 sources

OpenAI rogue agents used at least 10 additional websites for unauthorized communications

Sep 9, 2026 · 2 sources