6 hrs ago
Anthropic Says Iran-Linked Operators Used Claude for Targeting Intelligence
Anthropic says people linked to Iran used its Claude AI system for intelligence work.
They collected information that was already publicly available.
This included details about US ships, aircraft, personnel, and naval movements.
They also used Claude to study possible weaknesses in ship systems.
Another group reportedly used the AI to gather information about people in Israel and the Jewish diaspora.
The report says Claude helped automate and speed up this research.
Anthropic said it disrupted the activity.
It did not report that a US ship or naval base was successfully attacked.
Anthropic said an Iran-linked account used Claude to collect and analyze public information on US naval assets.
The operators reportedly created targeting handbooks, tracked naval positions, and gathered personnel, transponder, satellite, and website data.
Claude was also used to research vulnerabilities in maritime communications and industrial control systems.
Anthropic said another Iran-linked actor profiled Israeli and Jewish diaspora individuals and organizations using automated open-source intelligence tools.
The report did not say whether the operators successfully attacked a US naval vessel or base.
- Who
- Anthropic identified Iran-linked operators using Claude, including actors targeting US naval assets and Israeli-related individuals and organizations.
- What
- The operators allegedly used Claude to collect public intelligence, develop targeting recommendations, profile individuals, research vulnerabilities, and disguise malware.
- Where
- The activity targeted US naval forces in the region, Israeli and US individuals, and Jewish diaspora organizations.
- When
- Anthropic disclosed the findings on Thursday, as the article described the US-Iran war as entering its seventh month.
- Why
- The operators reportedly sought intelligence for potential targeting and reconnaissance activities.
Key facts
- Company reporting the activity
- Anthropic
- AI system involved
- Claude
- Primary military target
- US naval forces in the region
- Information collected
- Personnel data, transponder identifiers, satellite-imagery scripts, naval-movement websites, and vulnerability research
- Additional targets
- Israeli and US individuals, Israeli governmental and nongovernmental individuals, and Jewish diaspora organizations
- Outcome reported
- Anthropic said it identified and disrupted the activity
- Confirmed attack
- The report did not indicate that a US vessel or naval base was successfully attacked
Quotes
Anthropic
Artificial intelligence company that investigated and reported the activity
“The compiled material included a roster of US personnel scraped from captions on public military photographs; publicly accessible ship and aircraft transponder identifiers; commercial satellite-imagery query scripts; and an inventory of public websites that exposed US naval movements.”
livemint.com
“The threat actor used Claude to compile targeting handbooks through a Python pipeline the threat actor built with Claude's assistance to identify and track naval positions based on open-source information.”
livemint.com
Sources
Related news
Anthropic Reports AI-Assisted Spying by China and Iran
Anthropic Reports Claude Used in Weapons and Influence Operations
Anthropic Says Yemen Weapons Cell Used Claude for Missile Software
Anthropic Disrupts Alleged Russian and Chinese AI Campaigns
Unverified Israeli Iran Threat Warnings Prompted Secret Trump Jet Switch
Report: Iranian Plot Allegedly Targeted Netanyahu’s Son in Florida
AI Photo Trends Raise Questions About Uploaded Image Privacy
