A new official report by Anthropic, published in September 2026, reveals fresh details about how security, military, and government entities linked to the Islamic Republic of Iran allegedly misused artificial intelligence models.
📌 Key threats identified in the report:
🔹 Cognitive warfare and media manipulation: Operating media command centers—including one known as “Manjanigh” in Mashhad—rewriting Islamic Revolutionary Guard Corps bulletins, and falsely attributing tailored claims to prominent Western think tanks to manufacture credibility.
🔹 Surveillance and profiling of citizens:Developing a centralized security platform called “Arman,” distributing a malicious Firefox extension known as “Al-Najm Al-Thaqib” to uncover users’ identities, and creating phishing pages designed to collect Iranian national ID numbers.
🔹 Development of spyware: Building the powerful SECOMS64 malware, reportedly capable of keylogging Telegram Desktop, stealing Microsoft 365 account information, and creating phishing pages disguised as tools for bypassing internet censorship.
🔹 Military and urban surveillance: Collecting intelligence to support the targeting of US military vessels in the region, as well as designing systems for vehicle license-plate recognition and mobile-device identifier tracking.
anthropic.com/threat-intelli…
📌 Key threats identified in the report:
🔹 Cognitive warfare and media manipulation: Operating media command centers—including one known as “Manjanigh” in Mashhad—rewriting Islamic Revolutionary Guard Corps bulletins, and falsely attributing tailored claims to prominent Western think tanks to manufacture credibility.
🔹 Surveillance and profiling of citizens:Developing a centralized security platform called “Arman,” distributing a malicious Firefox extension known as “Al-Najm Al-Thaqib” to uncover users’ identities, and creating phishing pages designed to collect Iranian national ID numbers.
🔹 Development of spyware: Building the powerful SECOMS64 malware, reportedly capable of keylogging Telegram Desktop, stealing Microsoft 365 account information, and creating phishing pages disguised as tools for bypassing internet censorship.
🔹 Military and urban surveillance: Collecting intelligence to support the targeting of US military vessels in the region, as well as designing systems for vehicle license-plate recognition and mobile-device identifier tracking.
anthropic.com/threat-intelli…
www.anthropic.com Countering misuse of AI: September 2026 / Anthropic Case studies from threat actors disrupted between December 2025 and August 2026 across seven areas of harm, from cyber operations to biological misuse. 
