Executive Summary<\/strong><\/p>\n The Islamic Republic of Iran is aggressively pursuing Artificial Intelligence (AI) to strengthen its national development and security. Official directives, such as the \"<\/span>National AI Document<\/span><\/a>\" and the \"Judiciary Transformation Document,\" reveal an ambition to join the leading countries in the field. However, Iran\u2019s path is distinct from Western or Chinese models; in the absence of a robust private sector or global integration, national security has become the <\/span>de facto<\/span><\/i> driver of AI policy.<\/span><\/p>\n Our report finds that while the state struggles with a massive investment gap\u2014investing less than $50 million compared to regional rivals' billions\u2014it is aggressively implementing AI in the judicial and security sectors. The result is a dual-track development: a struggling commercial ecosystem starved of capital, contrasting with a rapidly advancing state apparatus designed for control. The integration of AI into the judiciary, exemplified by the \"Tenad\" system trained on 110 million rulings, risks baking historical biases into the legal system and removing human oversight from due process.<\/span><\/p>\n To understand the trajectory of AI in Iran, observers must distinguish between four distinct layers of state directives that often compete for dominance:<\/span><\/p>\n The Vision: <\/b>The National AI Document<\/i><\/b> (2024)<\/b> Initiated by the Supreme Council of the Cultural Revolution. It sets the high-level goal for Iran to rank among the top 10 AI nations by 2032 and become a regional \"research hub\".\u00a0<\/span><\/p>\n Mandate: <\/b>The Judicial Transformation Document<\/i><\/b> (Updated 2024)<\/b> A specific operational mandate for the Judiciary. It explicitly calls for \"maximum smartification\" of courts, automated case reviews, and reducing human intervention in expert referrals.<\/span><\/p>\n Legislation: <\/b>The National AI Bill<\/i><\/b> (Under Review)<\/b> Proposed legislation currently in Parliament to formalize the governance structure. It frequently conflicts with executive orders, creating a complex web of \"organizations\" vs. \"headquarters\".<\/span><\/p>\n Roadmap: <\/b>National AI Development Program<\/i><\/b> (August 2024)<\/b> A 23-page executive roadmap released by the \"Headquarters,\" detailing 15 axes of implementation until the final governance structure is determined.\u00a0<\/span><\/p>\n The roadmap for Iran's technological future is cluttered with conflicting mandates. While the <\/span>National AI Document<\/span><\/i> speaks of \"social justice,\" the operational reality is increasingly defined by security imperatives. Unlike Western or Chinese models where economic dominance is a primary driver, Iran\u2019s AI policy is viewed primarily through a security lens.<\/span><\/p>\n The <\/span>National AI Document<\/span><\/i> explicitly links technological advancement to \"enhancing governance quality\" and \"national security\". This approach treats individual rights\u2014such as privacy and free expression\u2014not as protected values, but potentially as obstacles to the regime's geopolitical ambitions. As noted by analysts at <\/span>Recorded Future<\/span><\/i>, Iran\u2019s state-centric model focuses heavily on control capabilities that could stifle innovation while amplifying the state\u2019s repressive capacity.<\/span><\/p>\n The governance of this strategic vision has been anything but stable. The administration of AI in Iran is characterized by rapid, often contradictory structural changes that suggest deep internal power struggles.<\/span><\/p>\n This volatility reached a peak in mid-2025. In June 2024, the Supreme Council approved the formation of a \"National AI Organization\" as an independent body under the President. Yet, by May 13, 2025, the government effectively shelved this organization, approving instead a \"Headquarters for the Development of AI Technology\" under the Vice Presidency for Science. Just four days later, Parliament approved the generalities of the <\/span>National AI Bill<\/span><\/i>, which would revive the National AI Organization, effectively overruling the government's previous move .<\/span><\/p>\n Until these hierarchies are clarified, the multitude of players\u2014including the Supreme Council, the Vice Presidency, and Parliament\u2014intensifies the risk of parallel efforts and wasted resources. The current executive roadmap serves as a temporary guide, but the lack of a unified command stifles long-term planning.<\/span><\/p>\n Iran\u2019s ambition to be a regional AI hub faces a harsh reality: international sanctions and economic isolation have created a funding chasm compared to its neighbors.<\/span><\/p>\n In April 2025, Communications Minister Sattar Hashemi highlighted the severity of the situation, noting that while Saudi Arabia invests approximately $20 billion in smartening and AI infrastructure, and the UAE invests $1.4 billion, Iran\u2019s investment stands at less than $50 million.<\/span><\/p>\n The internal budget allocation reflects this scarcity. The National Development Fund has allocated 5.6 trillion Tomans in loans to knowledge-based companies, with another 1.5 trillion planned for 2025. The Vice Presidency for Science has secured 4 trillion Tomans in the 2025 budget bill, and individual ministries have estimated budgets of merely 13-15 billion Tomans for the \"Government AI Assistant\" project. Critics, including the CEO of Zaeem Technology Development Group, argue that without specified resources, the National AI Document remains \"more aspirational than realistic\".<\/span><\/p>\n Beyond capital, the development of AI requires massive, stable energy supplies\u2014a resource in critical shortage in Iran. The <\/span>2025 IDCA Global AI Report<\/span><\/i> explicitly identifies power supply limitations as the main risk for Iran falling behind in the regional AI race. Developing countries typically have access to only 2% of the per capita electricity of developed nations, and Iran\u2019s aging power grid faces severe challenges in supporting the high-energy data centers required for training Large Language Models (LLMs).\u00a0<\/span><\/p>\n Furthermore, a critical failure of the strategic planning is the absence of a safety net for the workforce. The Iranian Parliament Research Center estimates that 20% of existing jobs in the Iranian labor market will be affected by AI in coming years. While the National Document aims for \"50,000 person-courses\" for employee training, this figure is negligible given the scale of potential displacement. Without broader educational reforms and accessible retraining schemes, the economic benefits of AI are likely to concentrate among a small elite, widening the gap between highly skilled professionals and the rest of the workforce.<\/span><\/p>\n Barred from accessing advanced global hardware due to sanctions, Iran is pursuing a dual strategy of \"technological self-sufficiency\" and cooperation with aligned non-Western powers.<\/span><\/p>\n The government is attempting to build an independent infrastructure stack. This includes the \"Sahand\" project, currently underway to produce necessary AI chips domestically, and the launch of GPU data centers, which reached the operational pilot phase at Pardis Technology Park in May 2025. These efforts build on earlier initiatives like Sharif University's \"High-Performance Computing (HPC) Center,\" piloted in 2013.<\/span><\/p>\n To bridge its technological gaps, Iran is also leveraging relationships with China and Russia. The \"25-Year Comprehensive Cooperation Program\" with China specifically mentions cooperation on smart technologies and AI, while similar agreements have been signed with Russia . However, experts warn that relying on these nations carries the risk of importing their surveillance-centric governance models. China and Russia widely use AI for state control, and their technical exports often come embedded with architectures of repression .<\/span><\/p>\n Domestically, the government relies on a network of trusted private companies to build its \"National AI Platform.\" Key players identified in the project include:<\/span><\/p>\n While 452 AI companies have been identified in Iran, the head of the Iranian AI Association notes this is negligible compared to 70,000 global companies. The sector grows at only 3-4% in Iran versus over 20% globally, highlighting the stifling effect of isolation.\u00a0<\/span><\/p>\n The most advanced operationalization of Iran's AI strategy is visible not in the commercial sector, but in the Judiciary. Under the guise of efficiency and \"Smart Justice,\" the state is building a vast automated legal infrastructure.<\/span><\/p>\n The Judiciary claims significant progress in digitization. Identifying a convict's assets now takes approximately four minutes via real-time database access, and more than 85% of prisoners participate in court hearings electronically. In the first ten months of 2024, nearly 2.7 million criminal record clearance certificates were issued, largely automatically.<\/span><\/p>\n However, the implications of this automation for human rights are profound. The flagship initiative is the \"Tenad\" system (Analysis and Review of Court Decisions), which utilizes a dataset of over 110 million judicial decisions to train algorithms. While intended to help judges identify \"conflicting rulings,\" this digitization creates a \"black box\" of justice.<\/span><\/p>\n Training AI on 110 million historical records from a system with documented due process issues guarantees that the algorithms will learn, replicate, and automate those same biases. If the AI\u2019s definition of a \"strong\" ruling\u2014a metric used by the Supreme Court to evaluate lower courts\u2014is based on harsh sentencing patterns or biased historical data, it will pressure judges to conform to those patterns to avoid negative evaluations.<\/span><\/p>\n Most concerning is the <\/span>Judicial Transformation Document's<\/span><\/i> explicit goal to \"eliminate human intervention\" in areas like expert referrals. This removal of human discretion threatens the nuance required for fair trials, particularly in Iran's unique legal context. Article 167 of Iran\u2019s Constitution and Article 3 of the Civil Procedure Code explicitly direct judges to refer to recognized Islamic sources and valid fatwas when laws are silent or ambiguous. This requires judicial reasoning and creativity\u2014capacities that current AI systems cannot replicate. An automated system, bound by historical data and rigid algorithms, risks violating these constitutional requirements for independent interpretation.<\/span><\/p>\n The intersection of AI and national security represents the most critical concern regarding Iran's development path. Reports from <\/span>Recorded Future<\/span><\/i><\/a> warn that Iran is likely increasing efforts to use AI for enforcing moral codes and monitoring dissent.<\/span><\/p>\n This surveillance capability is often introduced through \"functional creep.\" Tools like the \"Arbaeen AI Assistant\" or the \"Government AI Assistant\"\u2014budgeted at 15 billion Tomans per ministry\u2014normalize the collection of citizen data for public services. Once citizens are accustomed to these services, the underlying data collection can easily expand into surveillance. The Judiciary\u2019s push to provide \"online access to information databases of governmental authorities\" creates a centralized pool of citizen data accessible to the state, enabling \"predictive\" monitoring of protests or dissent .<\/span><\/p>\n Furthermore, the National AI Document emphasizes \"regulatory sandboxes\"\u2014environments where companies can test innovations with simplified rules. In a country lacking independent civil society oversight, these sandboxes risk becoming zones where high-risk surveillance tools, such as facial recognition, are operationalized without public scrutiny or ethical impact assessments.\u00a0<\/span><\/p>\n Ultimately, Iran\u2019s AI strategy reveals a state attempting to build a digital fortress on a foundation of economic instability. While the commercial ecosystem starves for capital\u2014investing mere millions where rivals invest billions\u2014the security apparatus continues to advance, fueled by a determination to achieve \"technological self-sufficiency\" at any cost. Projects like the \"Sahand\" chip initiative and the deepening integration with Chinese and Russian tech sectors are not merely economic endeavors; they are the blueprints for a closed-loop surveillance state designed to be immune to external sanctions and internal dissent.<\/span><\/p>\n Yet, this fixation on security may prove to be the regime's blind spot. By aggressively pursuing automation while ignoring the predicted displacement of 20% of the workforce, the state is ignoring the human cost of its own modernization. There are no safety nets for the millions of workers who stand to lose their livelihoods to the very algorithms the government is funding.<\/span><\/p>\n This creates a volatile irony at the heart of Iran\u2019s AI development. The government is rushing to build tools capable of predicting and suppressing unrest, yet its neglect of the economic fallout may inevitably fuel the very instability it seeks to control. In its race to digitize the state, the Islamic Republic risks engineering a society where surveillance is state-of-the-art, but the social contract has collapsed.<\/span><\/p>\n Read Part Two of the Report Here<\/a><\/p>\n","protected":false},"excerpt":{"rendered":" From National Ambitions to Surveillance Tools Executive Summary The Islamic Republic of Iran is aggressively pursuing Artificial Intelligence (AI) to strengthen its national development and security. Official directives, such as the \"National AI Document\" and the \"Judiciary Transformation Document,\" reveal an ambition to join the leading countries in the field. However, Iran\u2019s path is distinctContinue reading \"Control Over Innovation: Iran’s Paradoxical AI Development (Part 1)\"<\/span><\/a><\/p>\n","protected":false},"author":13,"featured_media":5649,"comment_status":"closed","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"om_disable_all_campaigns":false,"_monsterinsights_skip_tracking":false,"_monsterinsights_sitenote_active":false,"_monsterinsights_sitenote_note":"","_monsterinsights_sitenote_category":0,"footnotes":""},"categories":[3],"tags":[337,341,340,339,338,195],"class_list":["post-5647","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-investigations","tag-artificial-intelligence-development","tag-artificial-intelligence-in-iran","tag-information-control","tag-national-artificial-intelligence-document","tag-national-artificial-intelligence-organization","tag-surveillance","entry"],"acf":[],"aioseo_notices":[],"aioseo_head":"\n\t\t\n\t\n\t\n\t\n\t\n\t\n\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\tQuick Guide: The Hierarchy of State Documents<\/b><\/h4>\n
The Strategic Vision: Ambition vs. Reality<\/b><\/h4>\n
Institutional Friction and Policy Whiplash<\/b><\/h4>\n
The Economic Reality: Ambition in Isolation<\/b><\/h4>\n
The Energy and Labor Crisis<\/b><\/h4>\n
Infrastructure and the \"Look East\" Strategy<\/b><\/h4>\n
The \"Knowledge-Based\" Ecosystem<\/b><\/h4>\n
\n
The Judiciary: \"Smart Justice\" or Automated Bias?<\/b><\/h4>\n
Surveillance and Security Dimensions<\/b><\/h4>\n
The Architecture of Control<\/b><\/h4>\n