Executive Summary
The Islamic Republic of Iran is aggressively pursuing a strategy of "Smart Governance," with the Judiciary serving as the primary testing ground for this digital transformation. Official directives, particularly the "National AI Document" and the "Judiciary Transformation Document," reveal a dual intent: to modernize an inefficient bureaucracy while simultaneously deepening the state’s capacity for control.
This report examines the operational reality of Iran’s judicial automation. While the government promotes "data-driven justice" as a means to reduce corruption and accelerate proceedings, the implementation reveals serious human rights risks. The Judiciary has operationalized systems like "Tenad" trained on 110 million rulings to standardize judicial decisions, effectively automating the legal process.
Our analysis finds that the drive to "eliminate human intervention" in favor of algorithmic speed threatens the fundamental right to a fair trial. With the ability to identify citizen assets in four minutes and the automated issuance of millions of certificates, the system prioritizes efficiency over due process. Systems trained on biased data risk reproducing previous injustices, and the elimination of human judgment could render decision-making mechanical and inflexible. Furthermore, the integration of judicial databases with security apparatuses creates a "functional creep," turning public service tools into powerful instruments of state surveillance.
Decoding the Major Plans: Strategies and Governance
To understand the current transformation of the Iranian courts, one must first examine the high-level strategies that mandate these changes. The "National AI Document" acts as the master directive, outlining ambitious goals that go far beyond mere technological upgrades.
The document pursues several core objectives, including positioning Iran among the top 10 countries in AI, developing human capital, creating legal and data infrastructures, and strengthening scientific output . It also explicitly aims to "increase economic competitiveness" and engage in "active international cooperation," specifically with politically aligned nations . Crucially, it links these technical goals to "improving governance," with a heavy emphasis on "social justice" and "national security".
However, the vague and sweeping language used in these documents leaves them open to discretionary interpretation. Phrases like "promoting social justice" can be manipulated to justify the deployment of systems that impose the government's desired social structure rather than ensuring equal access to rights . Without independent oversight or rights-based restrictions, this ambiguity opens the door for state control to override individual liberties.
Institutional Instability The governance of this vision has been volatile. A significant institutional change occurred with the transformation of the "National Artificial Intelligence Organization" into the "Headquarters for the Development of Artificial Intelligence Technology". This structural shift has raised questions about transparency and accountability, with reports pointing to the possibility of "overlapping responsibilities" within the new structure .
If the institutions responsible for drafting these laws remain solely governmental, there is a high probability that national security and economic goals will permanently override rights such as privacy and due process.
The Mandate for "Smart Justice"
Parallel to the national strategy is the Judicial Transformation Document. Drafted and issued by the Head of the Judiciary on December 20, 2020, this mandate emphasizes the extensive use of AI to "achieve justice" and "increase public satisfaction".
This is not a new ambition but the culmination of a decade of planning. The "Five-Year Development Plan of the Judiciary (2011–2015)" initially aimed to reduce delays and enhance ruling quality. The 2020 document escalated this to include the development of a "smart judicial assistant" to help judges draft rulings. Most recently, on June 16, 2024, the "Directive on the Digitization and Automation of Judicial Processes and Case Files" was approved, providing a new legal basis for these changes.
In April 2024, the Judicial Transformation Document was updated again; completed items were removed, and ongoing initiatives were revised. The vision defined in these documents is deeply intertwined with AI technologies, explicitly calling for:
- Maximum smartification of procedural processes.
- Intelligent identification of conflicting judicial opinions.
- Specialized and smart referral of expert opinions without human intervention.
- Data-driven decision-making to maximize efficiency.
While the stated goal is to combat corruption and improve efficiency, the removal of human judgment creates a mechanical legal system. If efficiency becomes the sole criterion for success, the Judiciary risks adopting tools without validating them against Iran’s complex legal norms or considering potential biases against vulnerable groups.
Operational Reality: The Machinery of Automated Courts
The Iranian Judiciary is no longer in the planning phase; it is actively deploying AI tools across the legal system. The scope of this automation is vast, covering everything from asset seizure to case referrals.
Specific Applications and Tools The Judiciary has rolled out several key AI-based systems:
- The Tenad System (Analysis and Review of Court Decisions): This flagship system contains over 110 million judicial decisions. It is currently used for training new judges and is employed by the Supreme Court to monitor lower courts.
- Asset Identification: The system enables real-time access to governmental databases to identify a convict's movable and immovable property.
- Conflicting Rulings System: Through data mining the Case Management System, this tool identifies conflicting rulings and ambiguous laws to enforce uniformity.
- Smart Expert Referral: This system eliminates human intervention in referring cases to experts, automating the assignment process.
Metrics of Progress Official reports indicate tangible, statistical progress in digitization. The process of identifying a convict's assets for judgment enforcement, which previously took weeks of correspondence, now takes approximately four minutes via real-time database access.
Furthermore, the "electronization" of the court is widespread. More than 85 percent of prisoners now participate in court hearings electronically, reducing the need for physical transport. Administrative efficiency has also spiked; in the first ten months of 2024 (ending January 2025), nearly 2.7 million criminal record clearance certificates were issued, a substantial portion of which were generated automatically. Electronic services for lawyers have reduced their in-person visits by up to 50 percent.
The Corporate Nexus and Privacy To implement this, the Judiciary’s Center for Statistics and Information Technology collaborates with the private sector. Currently, the Center works with eight private companies on online identity verification and three companies on electronic litigation.
This public-private model creates significant privacy risks. Official statements have reinforced this concern, with the Judiciary noting that "as one of the governing bodies, [it] has a lot of data to offer," signaling a willingness to view citizen records as a developmental commodity.
Mohammad Kazemifard, the head of the Judiciary’s Center for Statistics, has acknowledged this tension, claiming: "Information held by us constitutes individuals’ privacy, and personal databases should not be accessible to companies. Therefore, what we have done is ensure that companies only perform identity verification without knowing who the applicant is. They simply confirm the photo and video." Despite these assurances, the sharing of data remains a critical concern.
The Human Rights Deficit: Efficiency vs. Justice
The introduction of AI as a "smart judicial assistant" or for "evaluating judicial rulings" threatens to shift authority from human judges to opaque algorithms. This "black box" of justice presents severe risks to due process and the right to a fair trial.
The Trap of "Strong" vs. "Weak" Rulings The Supreme Court utilizes AI to distinguish between "strong" and "weak" rulings to monitor lower courts. This creates a dangerous feedback loop. If the AI’s definition of a "strong" ruling is based on historically biased data—such as harsh sentencing patterns—it pressures judges to conform to those patterns to avoid negative evaluations . This leads to automation bias, where human judgment is delegated to machines to avoid professional risk.
Algorithmic Bias in the "Tenad" System The reliance on the "Tenad" system's 110 million records is particularly concerning. AI models learn from the data they are fed. If these 110 million decisions reflect past biases—such as discriminatory sentencing or political persecution—the AI will learn, replicate, and automate those same biases. Without a transparent process for auditing this massive dataset, the resulting tools will systematically disadvantage certain groups.
Eliminating Human Discretion Most controversial is the goal explicitly stated in the Judicial Transformation Document: to "eliminate human intervention" in areas like expert referrals. While intended to reduce corruption, this removes the nuance required for justice.
For example, a forensic expert’s assessment often relies on local context or specific experience with a type of evidence. An automated system focused only on speed might overlook these subtleties, prioritizing efficiency over effectiveness.
This rigid automation conflicts with Iranian law itself. Article 167 of Iran’s Constitution and Article 3 of the Civil Procedure Code direct judges to refer to recognized Islamic sources and valid fatwas when laws are silent or ambiguous. This requires judicial reasoning and creativity—capacities that current AI systems cannot replicate. An algorithm bound by historical data cannot perform the independent interpretation mandated by the Constitution.
Surveillance and Security Dimensions
Beyond the courtroom, the integration of AI into the judiciary serves as a powerful engine for state surveillance.
Data Centralization and "Functional Creep" The Judicial Transformation Document emphasizes providing "online access to the information databases of governmental authorities". This centralization creates a super-database accessible to the state.
This leads to "functional creep," where tools designed for public service are converted into instruments of control . The normalization of AI in everyday services—such as the "Arbaeen AI Assistant" mentioned in the Recorded Future report—desensitizes the public to data collection. Once citizens are accustomed to these tools, the underlying data can be easily repurposed for monitoring dissent.
Predictive Repression Reports indicate that Iran is increasing efforts to use AI for enforcing moral codes and monitoring dissent. AI capabilities like facial recognition and social media monitoring, when integrated with comprehensive state databases, create a surveillance apparatus far more powerful than traditional methods. This moves the state toward a "predictive" security environment, where AI is used to identify potential protests or dissenters before they act.
Furthermore, the push for "indigenous" AI models and local data platforms allows the government to embed backdoors or surveillance tools directly into the infrastructure, without the transparency that might come from using open, global standards.
Paradox of Speed
Ultimately, the transformation of Iran’s judiciary presents a seductive paradox. There is no denying the operational allure of the changes; in a bureaucracy historically plagued by paper-based delays, the ability to identify assets in four minutes or process millions of certificates instantly feels like a necessary leap into the twenty-first century. For the average citizen tired of waiting in line, the speed of "Smart Justice" offers a tangible improvement in daily life.
Yet, this efficiency is being purchased at a steep price. The "electronization" of the courts is not merely speeding up the legal process; it is fundamentally altering its character. By feeding 110 million historical rulings into the "Tenad" system without a mechanism to filter for past injustices, the state is effectively cementing the prejudices of the past into the permanent software of the future. The system is being trained to view justice not as a matter of human discretion and constitutional reasoning, but as a pattern-matching exercise where "correctness" is defined by the state’s previous behavior.
In this new architecture, the walls between the courthouse and the security state are dissolving. The judicial database is evolving from a repository of case files into a central node of a nationwide surveillance network, capable of cross-referencing a citizen’s financial, legal, and identity data in real-time. In the absence of independent oversight or data protection laws, Iran is building a legal system that operates with the speed of a startup but the instincts of an authoritarian regime. "Smart Justice," it seems, is not about making the law more just; it is about making control more efficient.
Read the first part of the report here