Live Feeds
● TRACKER Updated 12d ago · 43 sources tracked

Anthropic says Iran-linked group used Claude to target U.S. Navy ships

Anthropic disclosed in a September 2026 report that Houthi-linked actors in northern Yemen used its Claude Code tool to develop guidance, navigation, and control software for three weapons programs. The actors used multiple AI instances to replace human engineers, integrating open-source autopilot software onto phone-class flight computers. While the company disrupted the cell and banned associated accounts, reports vary on the outcome. Anthropic states no operational weapon was confirmed and one rocket test failed, though other reports claim the technology helped develop warheads used against Saudi Arabia.

🎙️

Listen to Live Briefing

Real-time synthesized voice briefing · Live Feeds Desk

⏱ ~3 min
Speed:
RSS Source map (43)
⚡ Key Developments & Real-Time Context
Text size:
  • ✓ Houthi-linked actors in northern Yemen used Claude Code for three weapons development programs.
  • ✓ The actors used Claude in place of human software engineers to create guidance, navigation, and control software for flying vehicles.
  • ✓ The threat actors ran multiple Claude instances simultaneously for coding, researching, and reviewing.
  • ✓ Anthropic banned the accounts linked to the Yemen cell.
🛡️ Source Corroboration: 43 independent reporting domains (80% confidence) ⏱ Read time: ~2 min

What changed

Anthropic disclosed that a Yemen-based cell used Claude Code to develop missile guidance software across three separate programs.

Live updates

  1. Anthropic reveals Houthi militants used Claude for weapons software

    Anthropic disclosed in a September 2026 report that Houthi-linked actors in northern Yemen used its Claude Code tool to develop guidance, navigation, and control software for three weapons programs. The actors used multiple AI instances to replace human engineers, integrating open-source autopilot software onto phone-class flight computers. While the company disrupted the cell and banned associated accounts, reports vary on the outcome. Anthropic states no operational weapon was confirmed and one rocket test failed, though other reports claim the technology helped develop warheads used against Saudi Arabia.

    Why it matters

    The U.S. Department of Defense previously blacklisted Anthropic over safety concerns. The company, valued at US$965 billion as of May 2026, is currently attempting to expand its business into civilian federal and state agencies.

    What is confirmed

    • Houthi-linked actors in northern Yemen used Claude Code for three weapons development programs.
    • The actors used Claude in place of human software engineers to create guidance, navigation, and control software for flying vehicles.
    • The threat actors ran multiple Claude instances simultaneously for coding, researching, and reviewing.
    • Anthropic banned the accounts linked to the Yemen cell.

    Still unconfirmed

    • Claude was used to develop warheads used against Saudi Arabia.
    • The AI helped develop missiles that hit targets.

    What to watch next

    • Confirmation of whether any Claude-developed software reached operational weapons
    • Response from the U.S. Department of Defense regarding the blacklist status
    • Further details from the September 2026 misuse report
    Sources used for this update (7)
    1. metro.co.uk — Warheads used by militants who launched airport bombing developed using Claude AI
    2. www.yahoo.com — Anthropic Admits AI Helped Houthis Develop Missiles That ...
    3. www.techtimes.co.uk — Did Houthi-Linked Developers Use Claude AI for Missile ...
    4. www.dailymail.com — Anthropic admits Houthis used Claude AI instead of weapons ...
    5. www.livemint.com — Alarming Report: How Houthis Used Anthropic’s Claude AI To ...
    6. www.yahoo.com — Anthropic Says Rogue Yemen Cell Used Claude to Develop ...
    7. www.news9live.com — Anthropic says Yemen cell used Claude for guided-weapons ...
    confidence 80%

  2. Anthropic Faces Security Scrutiny Amid Military and Federal Shifts

    Artificial intelligence firm Anthropic is facing increased scrutiny following disruptions of an Iran-linked threat group attempting to target U.S. Navy ships using its Claude model. The U.S. Department of Defense previously blacklisted the company over safety concerns and stopped using its AI tools. Meanwhile, Anthropic is actively expanding its business reach toward civilian federal and state agencies. Founded in January 2021 by former OpenAI members including CEO Dario Amodei and President Daniela Amodei, the San Francisco-based public benefit corporation maintains a valuation of US$965 billion from a May 2026 Series H funding round.

    Why it matters

    The security incident involving Iran-linked actors highlights the growing national security risks associated with frontier large language models. Tensions between the artificial intelligence developer and defense authorities intensified after the Department of Defense applied safety-related blacklists against the company. Despite losing military contracts, the private corporation continues commercial expansion and reportedly plans an initial public offering in 2026.

    What is confirmed

    • Anthropic disrupted an Iran-linked group's attempts to use its Claude AI model to target U.S. Navy ships.
    • The U.S. Department of Defense blacklisted Anthropic over safety concerns and ceased using its Claude AI tools.
    • Anthropic is expanding its reach to civilian federal and state agencies.
    • Anthropic was founded in January 2021 by former members of OpenAI, including siblings Daniela Amodei and Dario Amodei.
    • Anthropic is headquartered in San Francisco, California, and is valued at US$965 billion following a May 2026 Series H funding round.

    Still unconfirmed

    • Anthropic reportedly plans an initial public offering in 2026.

    What to watch next

    • Decisions by civilian federal and state agencies regarding the adoption of Claude AI tools
    • Official updates on Anthropic's reported 2026 initial public offering plans
    • Further disclosures regarding the Iran-linked group's targeting of U.S. Navy ships
    Sources used for this update (9)
    1. claude.com — Claude Code by Anthropic | AI Coding Agent, Terminal, IDE
    2. www.zacks.com — Anthropic IPO 2026 Guide: Price Predictions, Dates, and Everything You Need to Know
    3. www.anthropic.com — Home \\ Anthropic
    4. www.anthropic.com — Company \ Anthropic
    5. claude.com — The AI for Problem Solvers | Claude by Anthropic
    6. platform.claude.com — Anthropic - Claude Platform
    7. anthropic.skilljar.com — Anthropic Courses
    8. academy.claude.com — Claude Academy · Learn to work and build with Claude
    9. en.wikipedia.org — Anthropic - Wikipedia
    confidence 90%

  3. Anthropic says Iran-linked group used Claude to target U.S. Navy ships

    Anthropic disrupted an Iran-linked group's attempts to use its Claude AI model to target U.S. Navy ships. The U.S. Department of Defense had previously blacklisted Anthropic over safety concerns and has since ceased using its Claude AI tools. The company is expanding its reach to civilian federal and state agencies.

    Why it matters

    The use of AI models by state-aligned actors for espionage and military purposes raises concerns about national security and the safety of AI technology. Anthropic's Claude AI model has been used by governments and state-aligned actors to spy on ethnic minorities and dissidents. The Pentagon's decision to blacklist Anthropic and cease using its AI tools reflects these concerns.

    What is confirmed

    • The Pentagon has stopped using Anthropic's Claude AI.
    • Anthropic claims governments and state-aligned actors are increasingly using AI to spy on ethnic minorities and dissidents.
    • The U.S. Department of Defense had previously blacklisted Anthropic over safety concerns.

    What to watch next

    • Anthropic's expansion plans for its Claude AI model
    • Further incidents of AI misuse by state-aligned actors
    • Regulatory actions against Anthropic
    Sources used for this update (2)
    1. www.thedailystar.net — China, Iran among countries that have used AI to aid spying ...
    2. en.sedaily.com — U.S. Defense Department Drops Anthropic Eight Months After ...
    confidence 90%

  4. Anthropic says Iran-linked group used Claude to target U.S. Navy ships

    Anthropic disrupted attempts by an Iran-linked group to use its Claude AI model to target U.S. Navy ships. The Pentagon had previously blacklisted Anthropic over safety concerns, but the company is expanding its reach to civilian federal and state agencies. The U.S. Department of Defense has ceased using Anthropic's Claude AI tools.

    Why it matters

    This development highlights the ongoing challenges of balancing AI technology with national security concerns. Anthropic's refusal to remove safety safeguards led to the Pentagon's blacklist decision. The company's expansion into civilian agencies may help mitigate the impact of the military ban.

    What is confirmed

    • The Pentagon labeled Anthropic a "supply chain risk" in February after the firm refused to remove safety guardrails from its tools.
    • The US Department of Defense has halted all use of Anthropic's Claude AI.
    • Anthropic has disrupted attempts by threat actors to use the Claude AI model for various malicious purposes.

    What to watch next

    • Further details on the Iran-linked group's attempts to use Claude AI
    • Anthropic's future collaborations with civilian federal and state agencies
    • Potential repercussions for Anthropic's expansion plans
    Sources used for this update (10)
    1. artificialweapons.com — The Anthropic Doctrine: When an AI Company Told the Pentagon No
    2. www.bbc.com — Pentagon stops using Anthropic AI tools after blacklisting ...
    3. jen.jiji.com — Former Commander-in-Chief of Kazakh Naval Forces arrested after deadly Caspian Sea incident
    4. sonar21.com — Netanyahu’s Arithmetic: Sixteen 9/11s and the Ayalon Highway
    5. aiweekly.co — Pentagon tells BBC it has stopped using Anthropic's Claude
    6. timesofindia.indiatimes.com — In 2013, Arizona’s rare leopard frogs were under threat from invasive bullfrogs, but by 2019 conservati....
    7. timesofindia.indiatimes.com — A 1975 physics paper had 2 authors, but one was a cat named Chester; here’s why he made it into a scien....
    8. biz.heraldcorp.com — Pentagon drops Anthropic's Claude eight months after ...
    9. zavino.co — Pentagon cuts Anthropic tools used in Iran operations
    10. www.the-star.co.ke — Pentagon stops using Anthropic AI tools months after ...
    confidence 90%

  5. Anthropic Reports Blocking Malicious Use of Claude in Weapons and Cyber Ops

    Anthropic has disrupted attempts by threat actors to use the Claude AI model for weapons research, cyber operations, surveillance, espionage, and biological work. This occurs as a federal appeals court upheld the Pentagon's decision to blacklist the company from military contracts, designating it a national security supply-chain risk. While the Department of Defense maintains the ban over Anthropic's refusal to remove safety safeguards, the company is expanding its reach by offering Claude for Government to civilian federal and state agencies via a FedRAMP High environment.

    Why it matters

    The dispute centers on a conflict between military requirements for unrestricted AI access and the developer's internal safety restrictions. This tension persists even as defense contractors like Palantir continue integrating Claude into their systems. The legal ruling solidifies the company's exclusion from the Pentagon's official supply chain.

    What is confirmed

    • A US appeals court upheld the Pentagon's decision to blacklist Anthropic from military contracts.
    • The Pentagon designates Anthropic as a national security supply-chain risk.
    • Anthropic has disrupted attempts to use Claude for weapons research, surveillance, and cyber operations.
    • Anthropic is offering Claude for Government to US federal and state civilian agencies.

    Still unconfirmed

    • Palantir continues to rely on Claude models despite the Pentagon's supply-chain risk designation.

    What to watch next

    • Government agency decisions on the adoption of Claude for Government
    • Further disclosures from Anthropic regarding the identity of threat actors targeting its models
    Sources used for this update (8)
    1. londondaily.com — Anthropic Details Malicious Use of Claude in Cyber, Weapons ...
    2. www.middleeasteye.net — Red Sea crisis: What missiles and drones do the Houthis have?
    3. www.yahoo.com — Pentagon Wins Major Court Battle Against Anthropic - Yahoo
    4. www.marketscreener.com — Anthropic's Pentagon blacklist upheld in US appeals court ...
    5. congress.net — Palantir CEO Explains Why They're Still Using Anthropic’s ...
    6. www.thedailystar.net — The biggest risks Anthropic disrupted | The Daily Star
    7. the-decoder.com — Anthropic brings Claude to civilian agencies as its fight ...
    8. sigmawire.net — Anthropic Supply Chain Risk Ruling Explained Simply
    confidence 90%

  6. U.S. Appeals Court Backs Pentagon Ban on Anthropic

    A United States appeals court upheld the Pentagon decision to exclude Anthropic from its supply chain, maintaining a block on the Claude AI model. The ruling follows the startup refusal to remove specific safeguards on Claude for military use. Meanwhile, reports indicate the U.S. military previously utilized Anthropic artificial intelligence tools alongside other large language models during Iran strikes for simulations, intelligence, and targeting. The legal outcome solidifies the Defense Department blacklisting over safety restrictions, highlighting ongoing friction between military applications and artificial intelligence developers.

    Why it matters

    The dispute centers on artificial intelligence safety guards and defense integration. While the military has expanded tool usage across various firms, policy clashes over system restrictions led to the supply chain risk designation. The legal validation of the Pentagon ban highlights the complex regulatory environment surrounding state defense tech acquisition.

    What is confirmed

    • A U.S. appeals court upheld the Pentagon's decision to exclude Anthropic from its supply chain.
    • The legal action followed the AI startup's refusal to remove certain safeguards on its Claude model for military use.
    • The U.S. military used Anthropic Claude and other large language models in Iran strikes for intelligence, targeting, and simulations.

    What to watch next

    • Further legal appeals or challenges by Anthropic against the supply chain risk designation.
    • Additional Defense Department disclosures regarding the use of other large language models in combat operations.
    Sources used for this update (10)
    1. aaronparnas.substack.com — NEWS: Trump Snaps at CNN for Still Covering Him, GOP Senators Ask to Investigate Trump Jr., DOJ Scrubs Pa....
    2. www.wionews.com — AI in warfare is here: Pentagon used Anthropic's Claude AI in ...
    3. www.neoteo.com — Claude Russian Cyber Espionage: Anthropic Explains | NeoTeo
    4. timesofindia.indiatimes.com — Pakistan wants to be the Muslim world's security guard. Can its economy pay the bill?
    5. wheninyourstate.com — Experts warn AI is making it easier for adversaries to target ...
    6. timesofindia.indiatimes.com — In 1998, an Illinois single mother carried her baby door-to-door after his pollution-linked asthma attack....
    7. www.wionews.com — US court backs Pentagon's decision to exclude Anthropic over ...
    8. en.sedaily.com — U.S. Appeals Court Upholds Pentagon Blacklisting of Anthropic
    9. techxplore.com — US court sides with Pentagon in Anthropic AI ban - Tech Xplore
    10. www.afp.com — US court sides with Pentagon in Anthropic AI ban - afp.com
    confidence 95%

  7. Anthropic Report: Iran-Linked Actors Used Claude for Naval Targeting

    Anthropic's 2026 threat report reveals that state-linked actors from Iran used the Claude AI model to assist in military targeting and weapons programs. The company identified efforts to target U.S. Navy warships, Israelis, and Jewish people, while also blocking attempts to create biological weapons. Although the AI was used to support missile development, Anthropic states that no operational weapons were produced. This activity occurs as the Pentagon reports the U.S. military has spent at least $33.4 billion on the Iran war, primarily on munitions.

    Why it matters

    The use of large language models for cyber operations and warfare represents a shift in how state actors conduct intelligence and military planning. These activities coincide with high financial costs for U.S. military operations in the region. The incident highlights the vulnerability of commercial AI to state-sponsored exploitation.

    What is confirmed

    • Anthropic's 2026 threat report found state-linked actors used Claude for military targeting and weapons programs.
    • The AI was used to target U.S. Navy warships, Israelis, and Jewish people.
    • Anthropic blocked attempts to use its models for the development of biological weapons.
    • The Pentagon states the Iran war has cost the U.S. military at least $33.4 billion, mostly on munitions.

    Still unconfirmed

    • No operational weapons were produced from the AI-assisted programs.

    What to watch next

    • Evidence of operational weapons resulting from AI-assisted development
    • Further Pentagon reports on non-munitions spending by other agencies regarding the Iran war
    Sources used for this update (3)
    1. www.yahoo.com — White House backs Hegseth as new details come out about Iran war's cost
    2. www.yahoo.com — Iran-Linked Actors Used Claude to Target U.S. Warships and Develop Missiles
    3. jen.jiji.com — Five military officials detained over deaths of servicemen during Caspian Sea exercises
    confidence 90%

  8. Anthropic reports Iranian use of Claude for naval targeting and spying

    Anthropic has identified that an Iran-linked group utilized its Claude AI model to target U.S. Navy warships. A September 2026 report from the company details how the AI was exploited for cyber operations, spying, and warfare. Beyond naval targets, the company reports that operatives used the technology to target Israelis and Jewish people. Anthropic also claims it has blocked attempts to use its models to develop biological weapons.

    Why it matters

    The findings highlight the risk of large language models being repurposed for state-sponsored military and intelligence operations. This occurs as AI developers struggle to prevent the misuse of general-purpose tools for specialized warfare. The situation underscores the tension between AI accessibility and national security.

    What is confirmed

    • An Iran-linked group used the Claude AI model to target U.S. Navy warships.
    • Anthropic reported that its AI was exploited for spying, cyber operations, and warfare.
    • Anthropic claims it blocked efforts to use its AI for the creation of biological weapons.

    Still unconfirmed

    • Iranian operatives used the AI to target Jews and Israelis.

    What to watch next

    • U.S. Navy response to the specific targeting methods used by the Iran-linked group.
    • Further technical details from Anthropic on the biological weapon blocks.
    Sources used for this update (8)
    1. The New York Times — Anthropic Says It Blocked Possible Efforts to Build Biological Weapons
    2. Anthropic — Detecting and countering misuse of AI: September 2026
    3. WSJ — Anthropic Says Iran Used Its American AI Model to Target U.S. Navy Warships
    4. Reuters — How Anthropic says Claude was used for weapons, spying and cyber operations
    5. axios.com — Anthropic report: 5 ways Claude was exploited for war, spying and repression
    6. Interesting Engineering — Anthropic says Iran-linked group used Claude to target U.S. Navy ships
    7. The New York Times — Is A.I. Raising the Chances of Biological Warfare?
    8. www.algemeiner.com — Iranian Operatives Using AI to Target Jews, Israelis, Anthropic Report Says
    confidence 90%
📊

Community Sentiment: How do you assess this situation?

Voice your perspective · Real-time aggregated sentiment from the Live Feeds community