ChatGPT, Gemini and beyond: Iran is leveraging US-made AI.
According to a recent report by the Financial Times, Tehran’s military and cyber apparatus is increasingly using Western-made artificial intelligence tools, including ChatGpt developed by OpenAi and Google’s Gemini, to support cyber operations, intelligence gathering, and military research.
The report states that entities linked to Iran have used these AI systems to enhance phishing campaigns, generate malicious code, create convincing online profiles, and identify vulnerabilities in target networks.
Cybersecurity experts and tech companies are convinced of this, highlighting how the latest AI models have expanded the operational possibilities of state-backed Iranian groups, including research on drone technology, electronic warfare, and battlefield decision-making systems.
The FT also recalls that the United Arab Emirates have reported suffering over half a million cyberattacks daily, facilitated by ChatGpt, and that Israelis have received waves of phishing emails and SMS messages, some of which apparently invited them to collaborate with Iranian intelligence.
All the details.
AI SUPPORTING IRANIAN CYBER OPERATIONS
According to a cybersecurity analyst from a large company, who asked the British financial daily to remain anonymous, “We are seeing signs indicating consistent use of AI-based prompts.” The expert added that “This has undoubtedly helped improve their capabilities.” Thanks to these tools, Iran has reportedly managed to maintain digital pressure against Israel and the United States, despite the fragile truce in effect since early April, identifying adversary cyber vulnerabilities while simultaneously strengthening its own defenses.
PHISHING AND BEYOND
Experts emphasize that AI is also used to create credible digital identities and conduct social engineering campaigns. Iranian attacks often rely on tricking victims into clicking malicious links but may require weeks of online interactions under false identities to gain the targets’ trust.
“All of this happens automatically,” Gil Messing of the Israeli cybersecurity company Check Point told the FT. “They are using every tool at their disposal to accelerate their operations through artificial intelligence.”
“If you come from Tehran and try to pose as the human resources manager of a defense sector company, it’s a tough task to talk to someone for a month and seem like a person living in California,” said the cybersecurity analyst.
WHAT GOOGLE DISCOVERED
Shortly before the conflict began, at the end of February, the Mountain View tech giant identified the APT42 group, state-backed by Iran, using Google’s Gemini model for this very purpose. The company had already detected that Iranian hackers employed the chatbot more extensively than groups linked to North Korea, Russia, and China.
GEMINI STUDIED TO DISRUPT F-35S
Last year Google discovered that Iranian hackers use its chatbot much more intensively than their North Korean, Russian, or Chinese counterparts, despite those being more sophisticated. The APT42 group had even used it to study how to disrupt American F-35 fighter jets.
THE ROLE OF CHATGPT
Western tech companies have sought to limit the use of their tools by malicious actors. OpenAI stated it implements security measures to detect and block abusive activities, specifying that it regularly counters attempts to misuse its services by entities affiliated with Iran. “When we identify harmful activities, we take coercive actions, including disabling accounts, revoking access, or limiting functionalities used improperly,” OpenAI said in a statement, adding that it has security measures across its platform and that its most advanced models “are not widely available for use.”
THE INTEGRATION OF AI IN IRANIAN ARMED FORCES
However, Iran’s use of artificial intelligence is not limited to cyber operations. A Financial Times analysis of about 300 articles published over the past five years in Iranian military journals highlighted studies dedicated to the use of AI in electromagnetic warfare, battlefield decision support systems, drone guidance, and underwater targeting activities.
“Iran wants to keep pace with the most advanced armies,” Nicole Grajewski, an expert on Iranian military affairs at Sciences Po, explained to the British financial daily.
THE TECHNOLOGICAL RACE BETWEEN TEHRAN AND WASHINGTON
According to Farzin Nadimi, an expert at the Washington Institute for Near East Policy, Western tools like Google Translate had already in the past enabled the Iranian security apparatus to access international military research more quickly, overcoming the shortage of personnel with English and Russian language skills.
At the same time, the United States is also extensively using artificial intelligence in its military operations. The Pentagon reportedly employed Claude, Anthropic’s AI, the Wall Street Journal noted, in strikes against Iran. As Axios recalls, Anthropic’s AI model was used in the operation to capture Nicolás Maduro and could plausibly have been used in a potential military operation in Iran. Its artificial intelligence is integrated into the Maven Smart System, developed in collaboration with Palantir Technologies. The latter is the most controversial Californian data mining company, founded by Peter Thiel and led by Alex Karp, with deep ties to US intelligence and military agencies, on Israel’s side in the war against Hamas.
Iran continues to suffer a significant technological gap compared to the world’s major powers due to international sanctions that have penalized its economy and research system.
Despite this delay, Tehran has demonstrated remarkable adaptability, concludes the Financial Times. According to some experts, the country is achieving significant results on the digital propaganda front, thanks to a network of accounts close to the state apparatus that regularly spread viral content generated with artificial intelligence, often used to ridicule former US President Donald Trump.
