AI in Government

OverviewThe issuesResources
Gov tools
OverviewTax administrationPublic financial managementRegulatory design & deliveryCivil service reformPublic procurementAnti-corruption & public integrityCivic engagement & open governmentPublic servicesLaw enforcement & disaster managementJusticePolicy evaluationBehavioural science

Anti-corruption & public integrity

Governments around the world are increasingly moving beyond experimentation and beginning to integrate AI into core integrity and oversight functions to better detect, deter and prevent corruption. AI is helping oversight bodies analyse large volumes of financial, legal and administrative data — revealing fraud, conflicts of interest, and suspicious patterns that would otherwise go unnoticed. By transforming how institutions understand risk and process information, AI is opening new frontiers for public integrity..

The current state of play

AI is gradually becoming a strategic tool for actors responsible for promoting integrity and combating corruption — including anti-corruption agencies, supreme audit institutions, and internal oversight bodies. As AI capabilities continue to mature, governments are increasingly using AI to:

  • Detect fraud and anomalies using machine-learning and pattern recognition. Algorithms can scan entire populations of transactions — rather than samples — to identify unusual behaviour outliers, inconsistencies or suspicious patterns. This enhances the ability to detect fraud, corruption risks or misallocation of public funds at scale.
  • Improve knowledge management and document analysis. Named entity recognition, semantic search and large language models (LLMs) support the rapid search, summarisation and analysis of legislation,  reports, contracts and case files. This enables more efficient auditing, legal analysis, and investigations.
  • Anticipate corruption risks through predictive analytics. AI systems can help anticipate potential corruption and fraud risks and other issues, enabling the prioritisation of cases for further human examination. Even in institutions without a mandate for ex-ante action, predictive analytics enable risk-based planning and resource allocation, and more effective prioritisation.
  • Analyse complex relationships and networks. AI increasingly supports the analysis of beneficial ownership structures, procurement relationships and other interconnected datasets, helping identify hidden connections, conflicts of interest and potential corruption schemes that would be difficult to detect manually.

Examples from practice

  • Lithuania: AI-driven fraud and corruption risk assessment. As part of ongoing work with the OECD, the Special Investigation Service of Lithuania developed a machine-learning-supported tool that integrates multiple data sources to identify patterns associated with fraud and corruption risks related to EU funds. The model is designed to help officers prioritise cases for further examination while preserving professional judgement and institutional accountability.
  • Brazil: Generative AI for audits. ChatTCU, developed by Brazil’s Federal Court of Accounts, assists auditors by retrieving case information, interpreting regulations and summarising findings. It is now being adapted by institutions across Latin America.
  • European Union: Ownership anomaly detection. The DATACROS system analyses corporate ownership structures across 44 countries to flag corruption and money laundering risks. It supports both investigative work and public transparency.
  • United Kingdom: Fighting benefit fraud. The Department for Work & Pensions uses AI to detect inconsistencies in Universal Credit claims. The system is backed by fairness safeguards, including human oversight and random audits to avoid automation bias (i.e., over-reliance on AI).
  • Colombia: Forecasting corruption risk in contracts. The VigIA system uses AI to flag high-risk public contracts before violations occur, enabling the City of Bogotá to focus investigative resources more strategically.

Untapped potential and the way forward

AI offers integrity bodies new ways to anticipate and address corruption risks – from analysing contracts and legislation to detecting hidden networks of influence. As tools become more advanced, they support early warning systems and dynamic risk monitoring. Emerging agentic AI systems may further enhance these capabilities by orchestrating multi-step analytical tasks, while keeping humans responsible for oversight and final decisions.

However, technology alone will not transform integrity systems. To unlock this potential, governments must ensure AI use is transparent, ethical and accountable. this requires appropriate human oversight, sound data governance, and effective risk management throughout the AI lifecycle. Investing in organisational capabilities, practical guidance and cross-institutional collaboration will be key to scaling trusted solutions.

Learn more

Review a detailed section on AI in fighting corruption and promoting public integrity here.

Other relevant OECD work: