Anthropic AI Threat Report: Bioweapon and Cyber Espionage Attempts Blocked
According to a newly published threat intelligence report, AI company Anthropic has blocked malicious uses of its Claude models and international cyberattacks.
In a recent threat intelligence report published by Anthropic, it was announced that the misuse of Claude artificial intelligence models for biological weapon research, Russia-linked cyberespionage activities, and unauthorized data scraping attempts by Chinese rivals has been blocked.
Intervention in Bioweapon Research
In its new threat intelligence report covering the past eight months, Anthropic shared examples of the misuse of Claude models for harmful purposes with the public.
The company reported that in five different cases, including a researcher planning bird flu mutation experiments, it detected and blocked attempts to use the models in ways that could support the development of biological weapons.
Russia-Linked Cyber Espionage Activities
The report also detailed targeted attacks in Ukraine by Russia-linked cyber espionage groups.
It was determined that a Russia-linked hacker group, noted to be aligned with Midnight Blizzard, utilized artificial intelligence in phishing, hijacking hotel Wi-Fi networks, and WhatsApp account infiltration operations.
Activities and Violations of China-Based Laboratories
The artificial intelligence company announced that attacks originating from seven China-based laboratories had also been neutralized.
It was alleged that Alibaba carried out a massive illegal data distillation attack, while organizations such as Moonshot and DeepSeek routed live customer chats through Claude to collect model responses as training data for their own AI systems.
Conventional Weapon Software and Global Threats
According to the report published by Anthropic, actors operating from different geographies such as China, Russia, and Yemen were identified as using Claude models to develop software for conventional weapons.
The company emphasized that security and intelligence efforts carried out on a global scale to prevent the misuse of artificial intelligence technologies will be maintained with determination.