Loading story…
A coordinated influence operation, codenamed PRISONBREAK, was uncovered by Citizen Lab and Clemson University researchers, revealing the use of artificial intelligence to incite unrest in Iran. The operation involved a network of over 50 inauthentic accounts on the social media platform X, which were created in 2023 but became active primarily in 2025. These accounts systematically disseminated AI-generated imagery, deepfake videos, and fabricated news content designed to encourage Iranian citizens to revolt against the Islamic Republic. The campaign's activity was closely synchronized with real-world military actions, notably the Israeli Defense Forces' strikes against Iranian targets in June 2025. Researchers found that the operation mimicked legitimate news outlets to lend credibility to its disinformation, and some posts achieved tens of thousands of views, despite overall limited organic engagement. The PRISONBREAK network targeted large public communities on X and may have paid to promote its content, amplifying its reach within Iranian online spaces. After evaluating alternative explanations, Citizen Lab concluded that the most plausible scenario is that an unidentified Israeli government agency, or a closely supervised contractor, orchestrated the campaign. The operation's primary objective was to destabilize the Iranian regime by leveraging advanced AI tools to manipulate public opinion and sow discord. The campaign's timing and messaging were at least partially coordinated with Israeli military operations, suggesting a hybrid approach combining kinetic and information warfare. The use of AI-generated content marks a significant evolution in influence operations, enabling more convincing and scalable disinformation efforts. While the campaign's direct impact on Iranian society appears limited, its ability to reach large audiences and its sophisticated use of technology raise concerns about future state-sponsored influence operations. The PRISONBREAK case highlights the growing intersection of geopolitical conflict, artificial intelligence, and social media manipulation. Researchers emphasize the importance of monitoring such operations, as they represent a new frontier in psychological and information warfare. The findings underscore the need for robust detection and mitigation strategies to counter AI-enabled disinformation campaigns targeting vulnerable populations during periods of geopolitical tension.
Track how attackers are adapting to this technology.
3 events from the most recent confirmed update back to the earliest known activity.
By October 3, 2025, CyberScoop reported on Citizen Lab's findings that the Israeli government was likely responsible for the AI-generated disinformation campaign targeting Iran. This represented a public attribution of the operation to a likely state sponsor.
On October 2, 2025, Citizen Lab published its investigation into PRISONBREAK in multiple language editions, describing an AI-enabled influence operation targeting Iran. The report framed the activity as a coordinated disinformation campaign intended to undermine the Iranian government.
Citizen Lab reported that an AI-enabled influence operation dubbed PRISONBREAK was aimed at destabilizing and ultimately overthrowing the Iranian regime. The campaign used coordinated online influence tactics and AI-generated content directed at Iranian audiences.
Vulnerabilities, threat actors, malware, products, organizations, and breaches Mallory has linked to this story.
Follow how adversaries are adapting to this technology, and where it touches your stack today.
4 references tracked. Mallory keeps watching after this page renders.
cyberscoop.com
Open sourcecitizenlab.ca
Open sourcecitizenlab.ca
Open sourcecitizenlab.ca
Open sourceMap indicators from this story to your assets and identify affected systems in minutes.
Every observed campaign, victim, and pivot linked to actors named in this story.
Malware, exploits, and IOCs connected to the activity described here.
YARA, Sigma, and Snort rules deployed to your SIEM as soon as they’re published.
Get matching new stories delivered to your team as they break — not the next morning.
Ask questions about this story and take action on the answers.
Other stories Mallory is tracking that overlap on entities, sources, or topic labels.
Iranian state-linked operations are increasingly using artificial intelligence as a force multiplier across cyber intrusions, influence activity, and broader hybrid warfare. Reporting on activity through the first half of 2026 says Tehran applied generative AI and large language models to speed up spearphishing, reconnaissance, malware development, and social engineering, while preserving familiar tradecraft such as credential theft, wipers, and hack-and-leak operations. The activity aligns with broader threat intelligence tracking that shows adversaries integrating AI into established workflows through experimentation, distillation, and operational deployment rather than relying on wholly new attack methods. The reporting also ties Iran’s AI-enabled activity to established intrusion sets and regional operations, including campaigns attributed to **MuddyWater**, and warns that the technology is helping scale information operations, domestic repression, and potentially military support functions. Analysts say foreign support and technology transfer from **Russia** and **China** are contributing to this progress, raising concern that AI-enhanced Iranian campaigns will continue to target Western and regional critical infrastructure, vital industries, and public trust with faster, more efficient cyber and influence operations.
Jul 16, 2026
Researchers at **Acronis** reported a cyberespionage campaign—tracked as **CrescentHarvest**—targeting Farsi-speaking individuals sympathetic to Iran’s anti-government protests, including potential dissidents and supporters abroad. The operation uses social engineering lures packaged as malicious archive files containing authentic-looking protest media and a Farsi-language report that portrays the protests positively and repeats protest slogans, likely exploiting heightened demand for information amid Iranian internet blackouts. The archive includes files disguised as a video and an image that deploy a previously undocumented malware family dubbed **CRESCENTHARVEST**, assessed to function as both a **remote access trojan** and **information stealer**. Reported capabilities include command execution, keystroke logging, and theft of sensitive data such as saved credentials, browser history/cookies, and **Telegram** account information, along with checks for installed antivirus to modulate behavior. While no specific APT group was named, Acronis assessed the code, infrastructure, and victimology as consistent with an **Iran-aligned** actor; reporting also noted operational quirks (e.g., unused C2 endpoints and buggy User-Agent handling) that may indicate lower maturity or a rapidly assembled campaign tied to the protest-driven geopolitical moment.
Jun 29, 2026
The UK National Cyber Security Centre, FBI, and Netherlands AIVD have disclosed an Iranian state-linked campaign using **CHOSEN BRICK** Windows spyware against dissidents, activists, journalists, and other perceived opponents of the Iranian regime, including victims in the UK, United States, and Netherlands. Active since at least 2025, operators impersonate trusted contacts on WhatsApp and Telegram, cultivate relationships with targets, and deliver tailored malicious files masquerading as legitimate applications or documents; one reported lure used fabricated MRI scan results. CHOSEN BRICK establishes reboot persistence, weakens Microsoft Defender protections, and uses victim-specific Telegram bots for command and control. It can collect screenshots, microphone audio, emails, files, browser-accessible WhatsApp and Telegram data, and system information, while exfiltrating data through Telegram and cloud object stores and deploying additional malware. The agencies warned that stolen information has appeared on pro-Iranian leak sites, enabling profiling and harassment and potentially increasing targets' physical-safety risks; they released indicators and mitigation guidance.
Sep 17, 2026