AI in Government
Justice
AI is reshaping how justice systems operate — offering new ways to improve efficiency, accessibility and reinforce the capacity to deliver people-centred justice. Governments are using AI to streamline court procedures, manage growing caseloads, and improve access to legal information and services. While promising, the use of AI in justice raises particularly high stakes: if inadequately or improperly designed or used, AI can amplify or perpetuate existing societal disparities and justice gaps. Governance frameworks and guardrails, policy tools to assess risks and incidents, and remedies for those affected by AI systems are important to ensure the equitable and trustworthy use of AI in justice.
The current state of play
AI is supporting justice systems in several areas:
- Making internal operations more efficient. AI helps automate administrative tasks such as case classification, legal research and document processing. In Argentina, Brazil, Colombia, France, Greece, Slovenia and Spain, for instance, courts use AI to route cases, summarise rulings and manage records — cutting delays and freeing staff for complex legal work.
- Improving delivery of justice services. Virtual assistants and chatbots are helping people navigate justice systems. Portugal’s “Practical Guide to Justice” chatbot offers legal information in plain language, while Greece’s i-ACCESS supports child victims through a co-designed legal chatbot.
- Enabling predictions for a more proactive justice system. Predictive tools are used to assess risk and prevent harm — for example, Spain’s VioGén system predicts domestic violence risk levels to guide protective interventions. These tools must be carefully monitored to avoid skewed outcomes and ensure human control.
- Monitoring judicial performance and enhancing accountability. AI is also being used to monitor court performance. Brazil’s RAFA 2030 system links judicial decisions to the UN Sustainable Development Goals, while Israel uses AI to improve oversight within its public defence system.
- While uptake is growing, challenges remain — including data quality, regulatory gaps, algorithmic opacity and limited AI skills in the justice sector. Over-reliance on AI or the risk of “black box” decisions often act as disincentives for justice professionals to use AI systems with confidence.
Examples from practice
- Spain: AI for legal document processing and anonymisation. Spain’s Ministry of Justice developed a suite of generative AI tools for document classification, summarisation and anonymisation. The system helps automate repetitive tasks and protect personal data.
- Brazil: AI assistant MARIA for Supreme Court drafting. The MARIA system summarises documents and drafts reports to assist court clerks at Brazil’s Supreme Court. It helps identify precedents and cut down drafting time, allowing legal staff to focus on complex judgement work.
- France: AI-driven anonymisation for open release. The French Court of Cassation’s AI-powered anonymisation engine is a key component in its efforts to increase transparency and accessibility of legal information. This system identify and protect personal data, preparing court decisions to for public release as part of the national open data efforts in the country.
- Argentina: Generative AI for employment cases. Buenos Aires courts now use ChatGPT to classify and draft decisions in employment disputes. This shift from older AI tools has significantly reduced ruling preparation time and improved processing speed.
- Peru: AI for urgent domestic violence protection. The Amauta Pro system automates resolution drafting for protection measures in domestic violence cases — reducing resolution drafting times from 3 hours to 40 seconds and improving victim safety.
- Colombia: Streamlining legal actions to protect human rights. PretorIA system assists the Constitutional Court in managing the high volume of tutela cases, or legal actions filed to protect human rights. With over 600 000 custodianships (“tutelas”) received annually, PretorIA helps pre-select cases for review by detecting predefined legal criteria.
Untapped potential and the way forward
AI can support justice systems by improving consistency, efficiency and accessibility. It can assist legal professionals, speed up processes and help people navigate complex procedures. But trust and fairness are essential. To avoid reinforcing challenges with skewed data in AI systems, AI must be developed with strong governance, judicial oversight and ethical design. Transparent and explainable systems, redress mechanisms and adherence to human rights standards — such those from the OECD and Council of Europe — are key to ensuring AI strengthens people-centred justice.
Learn more
Review a detailed section on AI in justice administration and access to justice here.
Other relevant OECD work: