In its latest security report, the American company Anthropic announced that networks affiliated with the security agencies of the Islamic Republic have utilized artificial intelligence to identify, track, and build cases against government dissidents. According to the report, one such network monitored over 6,000 individuals in a single year and analyzed more than 155,000 pieces of content published on social media to identify individuals and examine their communications and activities. Dissidents of the Islamic Republic and Iranians living abroad were also among the targets of these operations.
The investigation by Anthropic indicates that these activities were not limited to monitoring social networks. Actors linked to Iran utilized Claude artificial intelligence to develop phishing tools and malware, identify anonymous users, and link phone numbers to the real identities of individuals. Some of these tools were capable of recording keystrokes, capturing screen images, extracting information stored in web browsers, and collecting data related to Telegram.
The significance of this report lies in its demonstration that artificial intelligence is becoming an integral part of the surveillance and intelligence apparatus of the Islamic Republic. This technology has been used not only to analyze information but also to locate individuals, determine their identities, and develop espionage tools. Furthermore, the findings from Anthropic indicate that the scope of this surveillance extends beyond the borders of Iran to encompass dissidents and Iranian nationals residing abroad.