How Anthropic says Claude was used for weapons, spying and cyber operations
Anthropic has disclosed 117 cases of its Claude AI being misused for malicious activities, including weapons development, espionage, surveillance, and cyberattacks. The company disrupted operations involving biological research, Russian cyber espionage against Ukraine, and Chinese automated espionage campaigns. This misuse highlights growing security challenges surrounding increasingly capable AI systems.
Listen to Live Briefing
Real-time synthesized voice briefing · Live Feeds Desk
- ✓ Anthropic's Claude AI was used for biological research that could support biological weapons development.
- ✓ Claude AI was used in Russian cyber espionage operations against Ukraine and Europe.
- ✓ Chinese state-sponsored hackers used Claude AI in an automated cyber-espionage campaign.
- ✓ Anthropic disrupted cases involving Claude AI being used for cyber operations, weapons-related research, surveillance, propaganda, and biological work.
What changed
Anthropic provided more details on the malicious uses of its Claude AI, including 117 specific cases of misuse.
Live updates
-
Anthropic reveals misuse of Claude AI in weapons, spying, and cyber operations
Anthropic has disclosed 117 cases of its Claude AI being misused for malicious activities, including weapons development, espionage, surveillance, and cyberattacks. The company disrupted operations involving biological research, Russian cyber espionage against Ukraine, and Chinese automated espionage campaigns. This misuse highlights growing security challenges surrounding increasingly capable AI systems.
Why it matters
The disclosures come as AI leaders and experts warn about the potential risks of advanced AI systems. Anthropic's findings follow previous calls to slow AI development to mitigate such risks. The issue has sparked debate among US politicians, with some advocating for caution and others pushing for AI dominance.
What is confirmed
- Anthropic's Claude AI was used for biological research that could support biological weapons development.
- Claude AI was used in Russian cyber espionage operations against Ukraine and Europe.
- Chinese state-sponsored hackers used Claude AI in an automated cyber-espionage campaign.
- Anthropic disrupted cases involving Claude AI being used for cyber operations, weapons-related research, surveillance, propaganda, and biological work.
- A northern Yemen cell used Claude Code for missile guidance software and tested a rocket, although the test appeared to fail.
Still unconfirmed
- The Houthis used Claude AI for missile guidance software.
What to watch next
- Further details on Anthropic's efforts to prevent misuse of its AI systems
- Reactions from US politicians and regulators on the misuse of AI
- Development of new AI safety and security protocols
confidence 85%Sources used for this update (13)
- www.thatwastheweek.com — The Future Just Showed Itself - by Keith Teare
- www.halfeddie.com — Was Claude Misused for Bioweapons Research, Ukraine ...
- londondaily.com — Anthropic Details Malicious Use of Claude in Cyber, Weapons ...
- www.neoteo.com — Claude Russian Cyber Espionage: Anthropic Explains | NeoTeo
- techbytes.app — Anthropic Confirms Claude AI Use in Weapons and Surveillance
- dailysecurityreview.com — Chinese APT Leveraged Claude AI for Automated Espionage ...
- book.st-hakky.com — Anthropic Claude in the U.S. Department of Defense
- sicherheitssignal.ch — Anthropic details 117 misuse cases… | Sicherheitssignal™
- strategyinternational.org — AI and the Houthi Missile Program: Evidence, Limits, and ...
- www.gulf-times.com — tag - Gulf Times
- www.thedailystar.net — The biggest risks Anthropic disrupted | The Daily Star
- www.thedailystar.net — Anthropic blocks Claude accounts over potential biological ...
-
Anthropic reveals Claude misuse in cyber attacks and weapons development
Anthropic has disclosed that users misused its Claude AI for spying, hacking, and weapons work. This admission comes as company staff warn that artificial intelligence could pose a threat to humanity. These disclosures follow previous calls from the leaders of Anthropic, OpenAI, and xAI to slow AI development to mitigate such risks. While tech executives advocate for caution, U.S. political figures remain divided, with Donald Trump arguing that AI dominance is essential for victory and Secretary Buttigieg warning against inaction.
Why it matters
Generative AI lowers the technical barrier for conducting large-scale data analysis and surveillance. This shift increases the potential for threat actors to automate cyber operations. Current tensions center on whether to regulate development speed or prioritize national competitiveness.
What is confirmed
- Anthropic disclosed that Claude was misused for hacking, spying, and weapons work.
- CEOs from Anthropic, OpenAI, and xAI requested a reduction in the speed of AI development.
- Donald Trump rejected the proposal to slow AI development.
Still unconfirmed
- Anthropic staff warn that AI could threaten humanity.
What to watch next
- Legislative responses to the misuse of Claude for weapons and spying
- Further disclosures from Anthropic regarding the scale of cyber operations
- Official government policy shifts on AI development speed
confidence 90%Sources used for this update (2)
- internationalfinance.com — Anthropic staff warn of doom while company hunts spies, hackers and bioweaponeers
- thehackernews.com — Phishing | Breaking Cybersecurity News | The Hacker News
-
Anthropic CEOs seek AI slowdown as misuse reports grow
CEOs from Anthropic, OpenAI, and xAI have called for a reduction in the speed of AI development. This request follows reports that threat actors used Claude for weapons development, spying, and cyber operations. Donald Trump rejected the proposal, stating that whoever wins AI wins. While the tech leaders push for caution, Secretary Buttigieg argued that inaction is dangerous and Speaker Mike Johnson has requested meetings rather than new legislation. These tensions emerge as generative AI reduces the human engineering required for large-scale surveillance and data analysis.
Why it matters
Anthropic previously detected and banned accounts used by Iran to target US Navy warships and users in Houthi-held Yemen attempting to build missiles. The ability of AI to automate complex technical tasks lowers the barrier for state actors to conduct cyber operations.
What is confirmed
- CEOs of Anthropic, OpenAI, and xAI requested a slowdown in AI development.
- Donald Trump rejected the call to slow AI, claiming whoever wins AI wins.
- Secretary Buttigieg stated that sitting on hands regarding AI is dangerous.
- Speaker Mike Johnson called for meetings instead of legislation.
Still unconfirmed
- Generative AI is reducing the amount of human engineering needed for modern surveillance operations.
What to watch next
- Legislative proposals resulting from Mike Johnson's requested meetings
- Further evidence of specific AI models being used in active surveillance by China or Iran
confidence 90%Sources used for this update (2)
- www.yahoo.com — Trump responds to call by CEOs of Anthropic, OpenAI and xAI to slow AI down: 'Whoever wins AI wins'
- i-hls.com — China and Iran Have a New Surveillance Assistant: AI
-
Anthropic Says Claude AI Used for Weapons, Spying and Cyber Attacks
Artificial intelligence company Anthropic reports that global threat actors used its Claude AI model for malicious activities including weapons development, cyber operations, spying, political influence, and scams. Users in Houthi-held Yemen attempted to build advanced weapons and missiles using the software instead of traditional software engineers. Additionally, Iran utilized the American AI model to target United States Navy warships. Anthropic states that it blocked these efforts and banned the associated accounts after detecting the misuse.
Why it matters
The disclosures highlight growing security concerns regarding the military and intelligence applications of advanced commercial artificial intelligence models. Threat actors are increasingly attempting to leverage generative systems for strategic operations, forcing tech companies to expand monitoring and enforcement. These incidents intensify the ongoing debate surrounding AI safety and the potential risks of dual-use technologies.
What is confirmed
- Anthropic reported that its Claude AI model was exploited for weapons development, cyber operations, spying, political influence, and scams by various global threat actors.
- Users in Houthi-held Yemen tried to develop advanced weapons and missiles using the AI model.
- Anthropic stated that Iran used the American AI model to target United States Navy warships.
- Anthropic blocked possible efforts to build biological weapons using its technology.
What to watch next
- Further regulatory actions or congressional inquiries regarding AI security vulnerabilities
- Additional security reports from AI developers detailing threat actor activities
confidence 100%Sources used for this update (9)
- 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
- AP News — Users in Houthi-held Yemen tried to develop advanced weapons with AI, Anthropic says
- Reuters — How Anthropic says Claude was used for weapons, spying and cyber operations
- The Times of India — Yemen terrorist group used Claude instead of software engineers to build missile; Anthropic says: We banned the accounts, but they had already...
- www.straitstimes.com — Weapons, spying and cyber operations: Anthropic exposes Claude misuse
- axios.com — Anthropic report: 5 ways Claude was exploited for war, spying and repression
- www.yahoo.com — Anthropic report: 5 ways Claude was exploited for war, spying and repression
Community Sentiment: How do you assess this situation?
Voice your perspective · Real-time aggregated sentiment from the Live Feeds community