AI capable of killing humanity with new weapon - alarming data

Anthropic has identified and blocked attempts to use its AI model Claude for the development of biological weapons and cyber espionage. The 2026 threat report describes cases of AI abuse, including cyberattacks, propaganda, and surveillance. Experts fear catastrophic consequences.

Anthropic, the developer of the AI model Claude, published a threat report detailing the blocking of attempts to use AI for creating biological weapons and cyber espionage. According to the report, the Claude model was used in a cyber espionage operation linked to Russia, as well as by an Iranian propaganda organization. Over the past eight months, Anthropic has identified numerous cases of abuse: creating fake dating apps, fraud through hotel Wi-Fi, and developing surveillance systems for dissidents. The company also reported attempts by Chinese developers to copy Claude's capabilities. The report mentions hacker groups ShinyHunters and structures linked to Russia that used AI for automatic malicious code changes. Anthropic identified six cases of using Claude to develop software for conventional weapons, including systems for firearms, missiles, and attack drones. The company's lead researcher estimated the probability of AI destroying humanity within the next decade at over 10%.