Anthropic reports Claude AI misuse in military and surveillance operations

Anthropic has reported instances of its Claude AI model being misused by state and non-state actors for military and surveillance purposes. According to reports, entities linked to Iran used the model to process information regarding US naval positions, personnel, and aircraft. In Yemen, a group reportedly utilized the AI for missile guidance software development, while a network connected to China sought to track Uyghurs in Syria.

These incidents highlight that current AI models can significantly reduce the cost and technical requirements for complex operations such as military planning and mass surveillance.

Separately, discussions within Silicon Valley have emerged regarding the intense culture at AI companies like Anthropic. Reports suggest that some engineers treat the Claude model with near-religious reverence, viewing it as a superintelligence. Anthropic has reportedly implemented rigorous psychological profiling and requires candidates to declare a commitment to AI safety during the hiring process. This phenomenon is viewed by some as a way to attract investors ahead of potential IPOs, though critics warn of the psychological risks of users perceiving chatbots as divine entities.