TLDR
- Anthropic blocked attempts to use its AI models for biological weapons research, cyberattacks, and propaganda campaigns between December 2025 and August 2026
- One case involved a request to help draft a grant application for gain-of-function research on the chikungunya virus
- State-sponsored groups, spyware vendors, and criminal hackers were among those who misused Claude models
- None of the misuse cases involved Anthropic’s newer Claude Fable or Mythos-class models, except one distillation attempt
- Anthropic has tightened safeguards on newer models and shared intelligence with government authorities and industry partners
Anthropic published a threat intelligence report on September 11, 2026, detailing how its AI models were misused for dangerous activities over the past eight months.
Breaking News: Anthropic said it halted potential plots by scientists who used its A.I. models to do research that could have helped develop biological weapons. https://t.co/7Q01GZV0xA
— The New York Times (@nytimes) September 10, 2026
The report covers misuse of Claude between December 2025 and August 2026. It is the company’s third such report since March 2025.
Biological Weapons and Dangerous Research
The most serious cases involved attempts to use Claude for research linked to biological weapons. Anthropic flagged five case studies in this category.
In one case, a user asked Claude to help write a grant application for gain-of-function research on the chikungunya virus. This type of research genetically alters organisms to give them new or enhanced properties.
The proposed research aimed to make the virus more transmissible and better at evading the immune system. Chikungunya causes severe pain and fever and is spread by mosquitoes.
Anthropic said this kind of research could support vaccine development but could also be used to make a pathogen more dangerous.
The company said its older models, including Claude Opus 4 and Claude Sonnet 4.5, were below the threshold where they could meaningfully help with dangerous biological research. Newer models are more capable, so Anthropic has added tighter restrictions.
Cyberattacks, Propaganda and Scams
The report also outlined cases involving cybercrime and influence operations. Hacking group ShinyHunters was named, as were China-based labs.
A group with links to Russia’s Midnight Blizzard allegedly used Claude to build a system that rewrote malware code automatically whenever it was detected by security tools.
Anthropic also identified nine influence operations originating in Russia, Iran, Turkey, and regions including the Gulf, South Asia, Africa and Europe. These involved hundreds of fake social media accounts pushing coordinated political messages.
Other misuse cases included fake dating apps, hotel Wi-Fi scams, and surveillance tools designed to track dissidents.
None of the cases involved Claude Fable or the Mythos-class models, except one attempt to extract and replicate a model’s capabilities without permission.
Anthropic said it blocked all the activities described in the report. The company said it has used its findings to update its safeguards and has shared information with government authorities and industry partners.
The report comes shortly after researcher Jacob Coxon resigned from Anthropic, citing concerns that AI companies are moving too fast toward superintelligence without adequate safety measures.
Anthropic said it hopes the report will help other AI developers spot similar threats and give governments a clearer picture of how these threats are developing.
Stop guessing and start investing with confidence. KnockoutStocks gives you the AI insights, market intelligence, and stock research you need to spot opportunities, cut through the noise, and make smarter investment decisions — all in one powerful platform.
Sign up today and get 50% OFF full access to our premium stock picks.
Simply use coupon code SPECIAL50 at checkout to claim your exclusive discount.







