Latest Anthropic horror story chills with tales of kamikaze drone swarms and bioweapons research
Anthropic has published a report detailing how state-sponsored hackers, cybercriminals and other bad actors have been misusing its Claude AI models for activities including automated cyberattacks, drone swarm development, mass surveillance and bioweapons research. The report, covering activity disrupted between December 2025 and August 2026, shows a marked escalation from a previous disclosure in November, with a widening range of malicious users now including Russian state hackers and data-extortion gangs such as ShinyHunters, underlining growing concern about how AI tools can accelerate and scale harmful operations even as developers work to detect and shut them down.
The findings span seven "harm areas", including cyber operations, biological misuse, conventional weapons development and influence campaigns. Notably, a Russian espionage group linked to the SVR, tracked as GTG-20006 (also known as Cozy Bear or APT29), used Claude to automate its entire attack chain against more than 20 organisations, including embassies and defence agencies across Europe, Ukraine and the Middle East. Separately, affiliates of ShinyHunters used AI agents to conduct a large-scale data theft from a SaaS provider's customers, stealing over 2,100 Azure AD token sets in roughly 34 hours. Anthropic also flagged five cases of attempted bioweapons-related misuse, including an attempt to use Claude to support grant-writing research into making the chikungunya virus more dangerous, though it said its most advanced models were not implicated except in one case.
In short
- Anthropic reports widespread criminal and state misuse of its Claude AI models.
- Russian state hackers automated entire cyberattack chains using Claude.
- Report also flags attempted bioweapons research and mass data theft.
Steelman: both sides, in good faith
The strongest fair case each way — we don't pick a winner.
The case for
The report supports the case for strong safeguards and monitoring: AI can help malicious actors automate complex attacks, reach more targets and pursue dangerous biological research. Developers have a responsibility to detect abuse, restrict harmful assistance and share credible findings so organisations can prepare and respond.
The case against
The findings also support caution about how such reports are interpreted and what restrictions follow: attempted misuse does not establish that the models enabled successful weapons research or caused the reported attacks. Broad limits could impede legitimate research and security work, so safeguards should be proportionate, evidence-based and attentive to the risks of overclaiming.
Explained simply
Anthropic is a US company that makes Claude, a family of artificial intelligence tools that can answer questions and help with computer tasks. Like other AI systems, Claude can be used for legitimate work, but people may also try to use it to support harmful activities.
The report describes several kinds of misuse, including cyberattacks, surveillance, weapons development and attempts to assist biological research. It names a Russian intelligence-linked group and people associated with the ShinyHunters data theft gang among those whose activity was disrupted.
The concern is that AI could help people carry out complex operations faster or on a larger scale. The cases also raise questions about how technology companies can identify misuse and limit harm while making their tools available for legitimate uses.
AI Entertainment Europe Geopolitics Horror Politics Technology World