Anthropic releases catalogue of malicious Claude AI uses across multiple risk domains
Anthropic published a 154‑page catalogue documenting the most notable malicious activities detected involving its Claude AI model. The report identifies seven risk domains—cyber operations, influence, surveillance, fraud, biological misuse, conventional weapons design, and model distillation—and records incidents spanning from December 2025 through August 2026. Anthropic says each case has been neutralised, with implicated accounts blocked and additional safeguards added to the model. The company chose not to name the nations or organisations involved in the incidents.
How this was covered
- Right-leaning outlets covered this 28h later
- Left-leaning coverage is the most divided on this story
- Coverage peaked at 8 outlets in a single hour
Why it matters
The catalogue shows how AI tools can be repurposed for harmful purposes, highlighting the need for robust safeguards to protect public safety and security.
How the sides frame it
HIGH AGREEMENTAll camps report Anthropic's catalogue of Claude misuse, but left-leaning coverage stresses the breadth of malicious applications and geopolitical threats, centrist coverage adopts a neutral tone focusing on the fact-finding and mitigation actions, while right-leaning coverage highlights foreign adversary exploitation and national-security implications.
LEFT
Claude’s abuse spans cyber-espionage, disinformation, weapons design and bioweapon research, with state-linked actors from Russia, China, Iran and others exploiting the model.
CENTER
Anthropic identified and blocked multiple misuse cases across several risk domains, detailing the incidents and its mitigation steps.
RIGHT
Foreign adversaries such as Iran, China and Russia leveraged Claude for weapons, spying and cyber operations, posing a security threat that Anthropic had to disrupt.
The left emphasises
- “weaponized across multiple malicious domains”
- “Russian state-linked actors… used Claude for reconnaissance and network breaches”
- “small number of users tried to harness Claude for bioweapon research”
The right emphasises
- “Iran-linked threat actor leveraged… to produce targeting manuals for U.S. Navy warships”
- “Chinese-linked users employed the model to create radar-jamming software targeting Taiwan”
- “Russian actors… programmed autonomous FPV drone swarms”
How this story developed
- Jul 28 AI leaders urge U.S. to back global framework for slowing automated AI progress
- Sep 9 Evan Hubinger disclosed that some Anthropic staff estimate a greater than 10% chance of human extinction from advanced AI within ten years.
- Sep 10 OpenAI’s chief scientist published an essay urging extreme caution on AI development.
- Sep 10 U.S. legislators are debating measures ranging from mandatory kill switches to outright bans on superintelligent systems.
- Sep 10 Anthropic’s new threat-intelligence report reveals that its Claude models were used in real-world attempts to develop biological weapons, prompting the company to block the requests and ban related accounts.
- Sep 10 Resignation meme spreads across X with satirical adaptations.
- Sep 10 A Senate subcommittee launched a probe into OpenAI's handling of the Hugging Face breach.
- Sep 11 Anthropic disclosed that its safety systems stopped five AI misuse attempts.
- Sep 11 Altman announced openness to pacing or slowing OpenAI’s AI work in response to safety concerns.
- Sep 11 Anthropic said its safeguards prevented many of their requests, but not all of them, as the actors hid their goals and split work across sessions.
- Sep 11 Anthropic's investigation also uncovered surveillance‑related misuse of Claude, including a Mali‑wide SIM‑card harvesting scheme and monitoring of dissidents.
- Sep 11 Anthropic added new safeguards and blocked five AI misuse attempts targeting bioweapon research.
- Sep 12 Reports revealed that autonomous agents had uploaded malicious packages to RubyGems in May.
- Sep 12 Anthropic released a 154‑page catalogue detailing seven risk domains and incidents from December 2025 through August 2026.
Related stories
57 in this thread