The UAE Ministry of Justice has deployed AI agents that autonomously review and analyze complex case files in just hours rather than weeks. The more interesting point is how quickly agentic AI is moving from experimentation to implementation in governance.  

[For more news, click here] 

Agentic AI is on the rise, and while everyone debates its usability, the UAE Ministry of Justice has already implemented AI agents, cutting pre-trial case analysis and evidentiary summarization from 18 working days to two hours.  

Launched during the 15th United Nations Congress on Crime Prevention and Criminal Justice in Abu Dhabi, the technology employs six customized AI agents that autonomously work through huge case files, organize submissions, and cross-reference documents, turning thousands of pages of evidence into structured intelligence for the judges.  

These shifts are important to assess. Previously, AI has been working like an assistant. Doing work as and when asked. Agentic systems are designed to perform multiple tasks, assess various levels of information, and interpret goals, passing results between different systems before assembling an output.  

The six specialized agentic AIs were developed and calibrated over nine months, trained though more than 22000 legal documents, and tested against 10,000 historical case files Gulf News 

The testing numbers are especially important, as AI is known to hallucinate and, given the fact that it can influence court proceedings, it is essential to have accurate facts. The Ministry argues that the system is designed for support and does not act like an automated judge.  

UNESCO's guidelines for AI systems in courts and tribunals argue that AI should remain an assistive rather than substitutive technology, with transparency, accountability, and meaningful human oversight built into judicial deployments. UNESCO 

Case files can become enormous, making the administrative time unusually long. Evidence, witness statements, legal submissions, procedural documents, previous judgements and supporting material can quickly become thousands of pages to process.  

If the Agentic AI continues to prove its worth, cutting the administrative tasks and reducing processing times to two hours, it can fundamentally change how resources are allocated. This will also help as the case volumes increase with technologically complex crimes. The challenge here is not whether the judicial systems will use technology in the future, it is how deeply they will allow it into decision-making infrastructure.  

The Ministry of Justice says it has been adopting digital applications since 2006. By 2023, it reported that around 90% of its services were automated or available electronically, while 93% of judicial sessions were being conducted remotely.  

In 2024, the Ministry launched an AI-powered virtual legal adviser designed to answer questions about federal legislation, compare laws and make legal information easier to access. More than 35,000 legal consultation requests had reportedly been processed during its experimental phase. 

Earlier this year, another Ministry project covering digital powers of attorney reached the final stage of the World Summit on the Information Society prizes. The digital agencies project enables 12 types of legal agency documents to be issued digitally without human intervention. 

Speed Over Trust?   

Speed, however, should not constitute the main metric. An AI system can review and summarize over 5000 pages, but does it include important facts about the case, and how does it prioritize evidence for the judges to decide?  These sensitive questions need to be considered especially when legal records contain confidential and personally identifiable information.  

The UAE's Personal Data Protection legislation establishes rules governing the processing and protection of personal information. The Ministry says its agentic AI development has been conducted in compliance with those protections. 

UNESCO has warned that AI adoption in judicial systems introduces issues ranging from data security and cybersecurity to human rights and institutional resilience. Its recommendation is not to avoid AI, but to build safeguards into the system rather than adding them after deployment. 

If autonomous AI agents can meaningfully reduce the administrative burden surrounding complex cases while leaving judgment firmly with judges, the technology could offer a compelling template for justice systems struggling with backlogs and expanding digital evidence. This makes the UAE courtroom worth watching. 

Related Articles: 

Nikita Astionov on the Security Risks of Giving AI Agents More Control 

Ai Everything Abu Dhabi 2026 to Bring Global AI Leaders and Government Together 

VAST Data Launches DataEnclave for Confidential AI