A Yemen-based threat group linked to the Houthi-controlled north of the country attempted to use Anthropic’s Claude AI to develop software for ballistic missile systems, according to an Anthropic intelligence report cited by the Financial Times.
Anthropic said its safeguards blocked many requests, but some attempts succeeded after users concealed their objectives and divided their work across separate sessions.
The disclosure comes as Houthi forces expand their military presence along Yemen’s Red Sea coastline and move closer to the strategic Bab al-Mandeb Strait, a key global shipping route.
Related Tweet:
Yemeni militants used US’ Anthropic AI in ballistic missile building attempt — FT
— RT (@RT_com) September 11, 2026
They tried using it to ‘develop guidance, control and navigation software’
Anthropic said it blocked ‘many of their requests, but not all’ as they hid goals and split work across multiple sessions pic.twitter.com/ery28oc1RD
Also Read:

