Anthropic warned that artificial intelligence is becoming operational infrastructure for cyberattacks, surveillance (monitoring people or communications), political influence operations (organized efforts to shape public opinion), fraud and other forms of malicious activity. Its threat intelligence report covers misuse of Claude between December 2025 and August 2026, involving suspected state-sponsored groups, financially motivated criminals, commercial surveillance vendors and political operators across seven areas, including cyber operations, surveillance, influence campaigns, scams, biological misuse and conventional weapons development. Cases included a surveillance platform developed by a consultant working with Mali's national security authorities, Iranian surveillance software and a malicious browser extension, and a China-based network that deployed more than 4,700 AI personas (software-driven identities) across more than 20 dating apps. Those profiles contacted at least 25,000 people and generated roughly 2.36 million messages over two weeks in April. Anthropic also identified nine influence operations from Russia, Iran, Turkey and other regions that targeted audiences across six continents, including a Malaysia-focused system covering all 222 parliamentary constituencies and more than 1,000 fake X accounts. The company said AI is reducing the labor, expertise and cost of malicious operations rather than independently choosing targets. Anthropic banned accounts linked to the cases, upgraded detection systems and shared information with AI companies and authorities where appropriate, while acknowledging that removing access to Claude does not necessarily eliminate the broader threat.