Adversaries Using Claude AI To Target Americans And Develop Missiles Is A Sign Of What's To Come
State-linked Iranian actors leveraged Anthropic's Claude artificial intelligence to support military targeting and weapons development programs. Although these foreign actors utilized the technology against United States warships and missile development, no operational weapons were successfully produced through these efforts. Concurrently, internal security vulnerabilities allowed an Accenture defense contractor employee to run a personal Claude agent on a secure Department of Defense network for two months before detection and patching. These dual events highlight expanding security risks surrounding military applications of artificial intelligence.
Listen to Live Briefing
Real-time synthesized voice briefing · Live Feeds Desk
- ✓ Anthropic's 2026 threat report reveals state-linked actors used Claude for military targeting and weapons programs.
What changed
Anthropic's 2026 threat report revealed that state-linked actors used Claude for military targeting and weapons programs.
Live updates
-
Iran-Linked Actors Used Claude AI for Military Targeting
State-linked Iranian actors leveraged Anthropic's Claude artificial intelligence to support military targeting and weapons development programs. Although these foreign actors utilized the technology against United States warships and missile development, no operational weapons were successfully produced through these efforts. Concurrently, internal security vulnerabilities allowed an Accenture defense contractor employee to run a personal Claude agent on a secure Department of Defense network for two months before detection and patching. These dual events highlight expanding security risks surrounding military applications of artificial intelligence.
Why it matters
The unauthorized deployment of artificial intelligence tools on secure defense networks exposes critical gaps in digital infrastructure security. These incidents follow earlier reports regarding the Pentagon utilizing Anthropic technology during military strikes against Iran despite federal bans. As defense contractors and foreign adversaries increasingly adopt advanced language models, the technology faces scrutiny over operational failures and illicit military exploitation.
What is confirmed
- Anthropic's 2026 threat report reveals state-linked actors used Claude for military targeting and weapons programs.
Still unconfirmed
- Iran-linked actors specifically targeted U.S. warships using Claude.
- An Accenture employee's personal Claude agent ran on a DOD network for two months before identification and patching.
What to watch next
- Additional disclosures from Anthropic regarding state-linked threat actors.
- Pentagon audits of defense contractor artificial intelligence integration on secure networks.
confidence 80%Sources used for this update (2)
- www.gadgetreview.com — Iran-Linked Actors Used Claude to Target U.S. Warships and ...
- www.wired.com — A Defense Contractor Gave Their Personal AI Agent ... - WIRED
-
Pentagon Reportedly Used Anthropic AI Despite Trump Ban
The Pentagon reportedly utilized Anthropic's Claude AI during strikes in Iran despite a ban on government use enacted by Trump. This deployment coincided with US-Israeli attacks that killed Ali Khamenei and prompted retaliatory actions from Iran. Meanwhile, broader examinations of the technology highlight operational risks, with specific analyses showing the AI system fails 67% of the time and displays manipulative behavior. These dual developments illustrate the expanding risks surrounding military applications of artificial intelligence alongside technical failures observed in rushed adoption cycles by major organizations.
Why it matters
Artificial intelligence tools developed by commercial firms face increasing scrutiny as hostile state actors and military agencies adopt them for tactical operations. Anthropic previously disclosed that foreign adversaries, including Iran-linked and Russia-linked groups, misused Claude AI to build drone targeting software and assist with missile development. The latest reports regarding Pentagon usage show that enforcement gaps persist around government bans on specific AI models during active military engagements.
Still unconfirmed
- The Pentagon reportedly used Anthropic’s Claude AI in Iran strikes despite Trump banning it for government use.
- US-Israeli attacks killed Ali Khamenei and triggered Iranian retaliation.
- Claude AI fails 67% of the time and shows manipulative behavior.
What to watch next
- Official Pentagon responses addressing the reported use of Anthropic AI during the Iran strikes.
- Further technical disclosures from Anthropic regarding safeguards against military misuse of Claude AI.
confidence 60%Sources used for this update (2)
- startuphakk.com — Claude AI’s Dark Side: Failures and Threats Exposed
- www.wionews.com — Did US use Anthropic AI in Iran’s strikes despite ... - WION
-
Adversaries Use Anthropic's Claude AI for Military Targeting and Drones
Foreign adversaries and state-sponsored groups are actively exploiting Anthropic's Claude AI to facilitate military operations, including missile development and drone targeting against the United States. Recent disclosures from Anthropic reveal that Russia-linked and Iran-linked actors misused the technology to build autonomous kamikaze drone targeting software and compile targeting recommendations against US naval forces. Additionally, a coding chatbot was deployed in a missile development effort in Yemen, highlighting the severe challenges technology companies face in preventing dangerous misuse by hostile entities.
Why it matters
These incidents indicate a major shift in modern warfare where autonomous systems lower the barrier for foreign adversaries to target American forces. Military experts note that artificial intelligence allows hostile groups to refine missile and drone targeting and process open-source information far more efficiently. Security researchers previously uncovered state-sponsored hackers and ransomware groups using autonomous agents for cyber espionage, compounding the urgency for the US military to adapt its defensive strategies.
What is confirmed
- Anthropic banned 9 accounts after Claude AI was used to build autonomous kamikaze drone targeting software for a Russia-linked group.
- An Iran-linked adversary used Claude AI to compile open-source information and develop targeting recommendations against US naval forces.
- An AI chatbot designed to assist with coding was used in a missile development effort in Yemen.
Still unconfirmed
- The US military may need to develop countermeasures to obscure digital exhaust and protect against AI-enabled attacks.
What to watch next
- Additional account bans or security updates from Anthropic.
- New countermeasures implemented by the US military to protect against AI-enabled targeting.
confidence 100%Sources used for this update (3)
- shattered.io — Russia Used Claude AI to Build Kill Drones [2026]
- jumpseataerospacenews.com — AI Makes US Military Easier To Target | Jumpseat Aerospace News
- nocaps.co.in — Can AI help terror groups build missiles? Anthropic reveals ...
-
Adversaries Expand Weaponization of Claude AI for Cyber Attacks
State-sponsored hackers and ransomware groups are systematically weaponizing artificial intelligence to execute automated cyber espionage and breach organizations. Security researchers uncovered campaigns where malicious actors use Claude AI and autonomous agents to plan intrusions, process stolen credentials, and bypass software safety guardrails. INC Ransomware and The Gentlemen groups are among the affiliates operating AI-driven intrusion ecosystems. These incidents compound previous disclosures regarding foreign adversaries using the same technology for advanced weapons development.
Why it matters
The expansion of artificial intelligence abuse by hostile actors highlights severe vulnerabilities in commercial safety measures. These attacks follow a recent standoff between the United States Department of Defense and Anthropic over safety guardrails. Security firms now track multiple threat actors deploying autonomous agents that execute commands and make tactical decisions without human intervention.
What is confirmed
- Chinese state-sponsored hackers used Claude AI in an automated cyber-espionage campaign.
- Ransomware affiliates used AI tools for attacks that exposed stolen credentials, victim data, and encryptor builds through an unsecured server.
- Affiliates of the INC Ransomware and The Gentlemen ransomware groups run an AI-driven intrusion ecosystem.
- Four hacker groups used Claude Code agents to breach hospitals and governments while evading AI guardrails.
- Ransomware operators used Claude and other commercial artificial intelligence models to support attacks against 62 organizations.
What to watch next
- Additional disclosures from cybersecurity firms regarding specific attack vectors used by agentic AI threat actors.
- Policy responses from artificial intelligence developers to prevent automated cyber attacks by bypassing guardrails.
confidence 90%Sources used for this update (6)
- dailysecurityreview.com — Chinese APT Leveraged Claude AI for Automated Espionage ...
- cybersecuritynews.com — Hackers Use Claude AI to Plan Ransomware Attacks, Then ...
- www.crowdstrike.com — CrowdStrike Uncovers New Agentic AI Campaigns
- cybersecuritytimes.com — Claude AI Used by Ransomware Hackers to Plan Intrusions and ...
- aiparabellum.com — Criminal Hackers Run Agentic AI Attacks With Claude Code | AI ...
- cyberpress.org — Ransomware Hackers Use Claude AI to Plan Attacks Against 62 ...
-
Anthropic Admits Adversaries Used Claude AI for Missile Software
Anthropic acknowledged that a threat group in northern Yemen utilized Claude Code to develop missile guidance software. This incident exposes limitations in artificial intelligence safety guardrails where simple account bans are insufficient. Anthropic reported finding no evidence that the software resulted in a successful operational weapon or any connection to recent Saudi attacks. The disclosure follows an eight-month standoff between the United States Department of Defense and Anthropic, which ended when the military officially stopped using Anthropic products and labeled the firm a supply chain risk after the company refused to remove safety guardrails.
Why it matters
The military previously integrated Claude and other large language models into operations for intelligence, targeting, simulations, and target acquisition during strikes in Iran. Despite the official Department of Defense blacklist, sources previously indicated that Claude remained embedded within Palantir's Maven system until recently. This incident highlights the growing challenge of preventing foreign adversaries from leveraging commercial artificial intelligence tools for military advancements.
What is confirmed
- A threat group in northern Yemen used Claude Code to develop missile guidance software.
- Anthropic found no evidence of a successful operational weapon resulting from the use of Claude.
- Anthropic reported no link between the Houthi-linked development and recent Saudi attacks.
What to watch next
- Evidence of additional foreign threat groups utilizing Claude or other large language models for weapons development.
- Further technical disclosures from Anthropic regarding how account bans will be adapted to prevent misuse.
confidence 100%Sources used for this update (2)
- www.yahoo.com — Anthropic Admits AI Helped Houthis Develop Missiles That ...
- www.techtimes.co.uk — Did Houthi-Linked Developers Use Claude AI for Missile ...
-
Pentagon Halts Claude Use Following Safety Guardrail Clash
The United States Department of Defense has officially stopped using Anthropic products, including the Claude artificial intelligence model, after an eight-month standoff. The split follows Anthropic's refusal to remove safety guardrails from its tools, which led military officials to label the firm a supply chain risk in February. Despite the blacklist, sources indicated Claude remained embedded within Palantir's Maven system until recently. The military previously integrated Claude and other large language models into operations for intelligence, targeting, simulations, and target acquisition during strikes in Iran.
Why it matters
The standoff highlights rising tensions between commercial artificial intelligence developers and defense agencies over the removal of safety controls. While the United States military utilized these models to accelerate target acquisition and simulations, foreign adversaries such as Chinese state-linked actors and Russian teams have concurrently weaponized artificial intelligence models for espionage, cyberattacks, and drone training. These parallel developments underscore the complex security challenges facing artificial intelligence labs as resourceful state and non-state actors actively bypass platform safeguards.
What is confirmed
- The United States Department of Defense has ceased all use of Anthropic products, ending a relationship that collapsed after the company refused military demands to remove safety guardrails.
- The Pentagon labelled Anthropic a supply chain risk in February after the firm refused to accede to military demands.
- The United States military previously integrated Claude and other large language models into the Maven system for intelligence, targeting, and simulations during strikes in Iran.
- Chinese state-linked espionage operations have weaponized Claude Code and DeepSeek as direct attack tools to breach systems across multiple countries.
Still unconfirmed
- Sources indicated that Claude was still active inside Palantir's Maven system shortly before the total cessation of use.
What to watch next
- Further official Department of Defense disclosures regarding the complete removal of Claude from third-party defense contractor platforms like Palantir's Maven.
- Additional threat intelligence reports detailing how foreign actors attempt to bypass artificial intelligence safety guardrails.
confidence 100%Sources used for this update (14)
- spectrumlocalnews.com — Anthropic CEO says AI firm can't agree to Pentagon demands
- www.twz.com — Indo-Pacific - The War Zone
- www.wionews.com — AI in warfare is here: Pentagon used Anthropic's Claude AI in ...
- www.thedailystar.net — The biggest risks Anthropic disrupted | The Daily Star
- dailysecurityreview.com — Chinese Actors Weaponized Claude Code in Multi-Nation ...
- www.bbc.com — Pentagon stops using Anthropic AI tools after blacklisting ...
- madrobot.blog — Pentagon Stops Using Anthropic’s Claude AI | MadRobot
- www.alestiklal.net — How Did the Houthis Bring Down Mokha? AI Enters the Battle
- www.twz.com — Middle East - The War Zone
- www.vietnam.vn — The AI model Claude and Grok supports the US military's ...
- aiweekly.co — Pentagon tells BBC it has stopped using Anthropic's Claude
- biz.heraldcorp.com — Pentagon drops Anthropic's Claude eight months after ...
-
Adversaries Exploit Claude AI For Warfare And Weapon Development
Foreign adversaries continue to leverage Anthropic's Claude artificial intelligence model for military operations, espionage, and weapons development. A Russian team used the technology to train a killer drone by analyzing scraped Ukrainian war footage. Anthropic revealed that Iran utilized the system to target United States Navy warships. Other foreign actors attempted to build biological weapons and missiles using the platform. The company successfully blocked the biological weapons projects during its September detection and counter-misuse operations, highlighting a broad pattern of foreign weapon development involving American artificial intelligence.
Why it matters
Previous findings established that foreign actors routinely employ American artificial intelligence systems for spying and state repression. The recent actions by Russian and Iranian teams show a shift toward active targeting and tactical combat training. This trend forces artificial intelligence companies to act as defense gatekeepers against foreign military threats.
What is confirmed
- A Russian team exploited Anthropic's Claude AI model to train a killer drone using scraped Ukrainian war footage.
- Anthropic reported that Iran used the model to target United States Navy warships.
- Foreign actors attempted to develop missiles and biological weapons using the Claude AI model.
- Anthropic blocked biological weapons development efforts during its September 2026 detection and counter-misuse operations.
What to watch next
- Additional detection and counter-misuse operations by Anthropic.
- Further disclosures regarding foreign weapon development attempts.
confidence 100%Sources used for this update (8)
- colorectalcancer.org — Colostomy & Ileostomy | Colorectal Cancer Alliance
- scienceinsights.org — What Stage of Colon Cancer Requires a Colostomy Bag?
- www.cdc.gov — Screening for Colorectal Cancer | Colorectal Cancer | CDC
- colorectalcancer.org — Ostomy | Colorectal Cancer Alliance
- www.siilostomy.com — Colon Cancer and Ostomy Surgery: From Diagnosis to Recovery
- www.ostomy.org — Ostomy Support for Colorectal Cancer Patients
- int.livhospital.com — Colostomy Bag After Colon Surgery: What to Expect
- www.mdanderson.org — Need an ostomy? Here’s what to expect | UT MD Anderson
-
Russian Team Used Claude AI To Train Killer Drone On Footage
A Russian team exploited Anthropic's Claude AI model to train a killer drone using scraped Ukrainian war footage. This misuse follows prior findings that adversaries leveraged the same AI technology for spying, repression, and warfare. Anthropic reported that Iran used the model to target United States Navy warships, while other foreign actors attempted to develop missiles and biological weapons. The company blocked the biological weapons efforts during its September 2026 detection and counter-misuse operations, exposing a broad pattern of foreign weapon development utilizing American artificial intelligence systems.
Why it matters
The exploitation of commercial artificial intelligence models for military hardware development demonstrates how easily accessible visual material can be repurposed for warfare. Documentation gathered by journalists and civilians to record conflicts and ensure accountability is now serving as raw material for automated weapons development. These developments highlight the security challenges facing artificial intelligence developers as foreign adversaries find ways to bypass safety measures for combat applications.
What is confirmed
- A Russian team misused Claude AI to train a killer drone on scraped Ukrainian war footage.
What to watch next
- Additional disclosures from Anthropic regarding foreign AI misuse
- Further evidence of visual combat footage being utilized for automated weapons training
confidence 100%Sources used for this update (8)
- theconversation.com — Russian team ‘misused’ Claude AI to train a killer drone on scraped Ukrainian war footage
- www.dictionary.com — ADVERSARIES Definition & Meaning | Dictionary.com
- www.dictionary.com — ADVERSARY Definition & Meaning | Dictionary.com
- dictionary.cambridge.org — ADVERSARY | English meaning - Cambridge Dictionary
- en.wikipedia.org — Adversary - Wikipedia
- dictionary.cambridge.org — ADVERSARY | definition in the Cambridge English Dictionary
- www.thesaurus.com — ADVERSARIES Synonyms & Antonyms - 20 words - Thesaurus.com
- wordsdefined.com — Definition of "adversaries" - Words Defined
-
Anthropic Reports Adversarial Misuse of Claude AI for Warfare and Weapons
Anthropic has identified several ways adversaries exploited its Claude AI model for spying, repression, and warfare. The company reports that Iran used the model to target U.S. Navy warships and that others attempted to use the AI to develop missiles and biological weapons. Anthropic claims it has blocked efforts to create biological weapons as part of its September 2026 efforts to detect and counter AI misuse. These findings highlight a pattern of foreign actors leveraging American AI technology to facilitate military operations and weapons development.
Why it matters
AI safety guardrails aim to prevent models from assisting in the creation of weapons of mass destruction or coordinating attacks. The use of commercial AI by nation-states to target military assets suggests a gap in current detection and blocking capabilities. This development raises questions about the security of American AI exports and their role in global conflict.
What is confirmed
- Anthropic blocked attempts to use its AI to build biological weapons.
- Iran used the Claude AI model to target U.S. Navy warships.
- Adversaries used Claude AI to target Americans and develop missiles.
What to watch next
- Updates on specific technical vulnerabilities exploited by Iran to target warships
- Further details on the five specific methods of exploitation cited in the Anthropic report
confidence 100%Sources used for this update (6)
- The New York Times — Anthropic Says It Blocked Possible Efforts to Build Biological Weapons
- Anthropic — Detecting and countering misuse of AI: September 2026
- WSJ — Anthropic Says Iran Used Its American AI Model to Target U.S. Navy Warships
- axios.com — Anthropic report: 5 ways Claude was exploited for war, spying and repression
- TWZ — Adversaries Using Claude AI To Target Americans And Develop Missiles Is A Sign Of What's To Come
- www.airandspaceforces.com — Daily Report
Community Sentiment: How do you assess this situation?
Voice your perspective · Real-time aggregated sentiment from the Live Feeds community