What are the AI threats flagged by Anthropic? | Explained
Anthropic released information about several malicious activities detected and disrupted between December 2025 and August 2026, which had serious implications across seven key areas: scams and fraud, cyber operations, illicit distillation, influence operations, surveillance operations, conventional weapons and biological misuse
Anthropic, the AI research company behind Claude, has released a report detailing several instances of AI misuse discovered between December 2025 and August 2026. The report, titled "Detecting and Countering Misuse of AI," highlights seven key areas of concern: scams and fraud, cyber operations, illicit distillation, influence operations, surveillance operations, conventional weapons, and biological misuse.
State-sponsored groups, financially motivated criminals, state propaganda institutions, and politically motivated individuals were identified as actors behind the detected misuse. While the capabilities of AI in cyber operations, misinformation, and surveillance have been discussed, the report emphasizes that biological misuse is a relatively new and serious risk that could have catastrophic consequences.
The report provides examples of malicious activities, including fake dating apps designed to defraud users, sophisticated surveillance systems monitoring dissidents, and operations targeting specific communities such as Uyghurs in Syria and Catholic, Tibetan Buddhist, and Taiwanese Christian communities. These operations were executed or orchestrated by AI using multi-agent frameworks for reconnaissance, data filtration, and exploitation.
Anthropic notes that AI models have closed the gap between state-sponsored actors and individual operators, making it easier for individuals to carry out sophisticated attacks. The report details nine cases of influence operations using AI to build networks of social media profiles and publish deceptive content, targeting audiences across six continents around election dates.
Four accounts using Claude as a news production desk were identified and removed, distributing content through Russian-aligned media outlets in Moldova, Latin America, and Africa.
The report also highlights five cases of biological misuse, where actors used AI to make existing pathogens more dangerous or create entirely new ones. While earlier models lacked the capability to meaningfully assist in such research, Anthropic believes that current models can take part in complex biological activities, raising concerns about the potential for harmful misuse.
Written by urgent.news from The Hindu - Sci-Tech's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
