This report, titled **“Detecting and countering misuse of AI: September 2026,”** summarizes how Anthropic’s Threat Intelligence team detected and disrupted malicious attempts to use Claude over an eight-month period. The activity covered **December 2025 through August 2026** and spans **seven harm areas**: cyber operations, influence operations, surveillance, scams and fraud, biological misuse, conventional weapons development, and illicit distillation.
The report says the cases are not representative of ordinary misuse, but instead highlight the **most notable and novel threats** observed so far. The actors involved included **suspected state-sponsored groups, financially motivated criminals, commercial spyware vendors, state propaganda entities, and politically motivated individuals**. Examples ranged from **fake dating-app fraud networks** to **surveillance systems used to identify dissidents**.
According to the report, the misuse involved **Claude Haiku, Sonnet, and Opus** models. It notes that **Claude Fable and Mythos-class models were not involved**, except in one illicit distillation case. The company says it **disrupted the operations**, used the findings to **improve safeguards**, and **shared intelligence with authorities and industry partners when appropriate**.
The report emphasizes that as AI models become more capable, the risk of misuse will grow unless developers and broader society strengthen defenses. It aims to help other AI developers spot similar abuse patterns and give governments and civil society a clearer understanding of emerging threats.
Article timeline
Build this article’s timeline
Extract dated events, key actors, and open questions using the same timeline analysis available in Signal.
This report, titled **“Detecting and countering misuse of AI: September 2026,”** summarizes how Anthropic’s Threat Intelligence team detected and disrupted malicious attempts to use Claude over an eight-month period.
🤖 AI prefill
📅When
December 2025 – August 2026
🤖 AI prefill
📍WhereWhere did this occur?
💡Why
Centers on misuse of AI, Threat Intelligence team, and malicious activity
🤖 AI prefill
Login to answer and earn XP for each gap you fill.
Manipulation Score: LowHow much this article uses emotional framing vs. straight reporting
Loaded Language1225%
Absolutist1015%
Threat/Panic2225%
Us vs Them815%
Engagement Bait1820%
Manipulation score look right?
Factual Claims (10)
mediumtechnology
Over the past eight months, the Threat Intelligence team identified and disrupted operations where threat actors tried to use Claude for malicious activity.
lowother
The report covers activity disrupted between December 2025 and August 2026.
mediumtechnology
The report covers seven harm areas: cyber operations, influence operations, surveillance, scams and fraud, biological misuse, conventional weapons development, and distillation.
mediumtechnology
Claude Haiku, Sonnet, and Opus models were used in the misuse cases.
mediumtechnology
None of the misuse cases involved Claude Fable or Mythos-class models, except one illicit distillation case.
mediumother
The cases described are the most notable and novel threat activity identified to date.
mediumgeopolitical
Threat actors included suspected state-sponsored groups, financially motivated criminals, commercial spyware vendors, state propaganda institutions, and politically motivated individuals.
mediumsocial
One example involved a network of fake dating apps designed to defraud users.
highsocial
Another example involved surveillance systems built to identify and monitor dissidents.
mediumother
The company says it disrupted the activity, strengthened safeguards, and shared intelligence with authorities and industry partners where appropriate.
Comments
Login to view and post comments.