Skip to main content

Artificial Intelligence at GAO

Jump To:

Overview

Our reputation for objectivity, accuracy, and nonpartisanship is the foundation on which all our work rests. It is what Congress expects and relies on—they can trust what we say. Our reputation has been earned over a century of fact-based, nonpartisan work. We will only pursue AI use that enhances our support to Congress.

Strategy

Our AI strategy seeks to integrate AI into our audit and business processes with the singular goal of enhancing our support of Congress in carrying out its constitutional responsibilities for the benefit of the American taxpayer. We will achieve this by using AI to enhance staff productivity, accelerate mission delivery, and drive operational excellence. 

View the AI Strategy (PDF, 15 pages)

AI Guardrails 

As GAO expands its use of generative AI, the strategy must establish clear guardrails to ensure AI-assisted outputs remain accurate, defensible, and appropriately scoped. Given GAO’s mission depends entirely on the credibility of its findings, AI-generated content that informs engagement work must be both human-reviewed before use and grounded in verifiable source material — not simply produced and accepted at face value

Use Cases

As of March 2026

This table provides an overview of current AI use cases at GAO and will be updated periodically.

Use CaseUse Case DescriptionPotential BenefitsMaturity PhaseRelevant Techniques
Federal Audit Clearinghouse Exploration ToolSearches for and sorts Single Audit data and findings from the Federal Audit Clearinghouse
  • Allows GAO auditors to easily analyze Single Audit data and efficiently access and summarize findings
  • Identifies critical and persistent issues in the Single Audits and facilitates data reliability assessments
  • Predicts risk areas
Operational
  • Machine learning
  • Large language model
  • Forecasting
Topic ModelingOrganizes large volumes of text, such as public comments from Regulations.gov
  • Groups contents by similar themes using keywords or phrases
  • Facilitates further analyses such as auditing of regulatory review processes 
Operational
  • Natural language processing
  • Topic modeling
Computer-Readable FormattingExtracts specific information from large volumes of documents
  • Increases efficiency and reduces manual document review
  • Formats information from a non-computer readable document (e.g., PDF) into a tabular format
  • Allows computer-aided analysis of the data
Operational
  • Computer vision
  • Machine learning
Legislative MandatesDetects and summarizes GAO legislative mandates
  • Increases efficiency and reduces manual document review
  • Identifies any legislation that includes a mandate calling for GAO to conduct an audit or review
  • Produces an automated summary of the context surrounding the identified mandate
  • Highlights potentially fragmented and duplicative mandates
Operational
  • Natural language processing
  • Large language model
  • Semantic matching
  • Regular expressions
GAO Employee Experience SurveySummarizes qualitative responses from annual GAO Employee Experience Survey
  • Identifies trends, patterns, and sentiments quickly
  • Improves survey interpretation with less manual intervention
  • Assists with root-cause analyses
Operational
  • Natural language processing
  • Large language model
Generative AI Tool for GAO StaffProvides Generative AI functionality to support multiple operational and audit efficiencies
  • A Generative AI system tailored to complement GAO's unique mission and operational needs
  • Improves knowledge management by providing a central access point for data sources critical to audit findings
Late-stage prototype
  • Large language model 
Congressional Outreach

Optimizes GAO delivery of products and information to members of Congress based on their interest areas

 

 

  • Improves GAO’s awareness of Congressional priorities
  • Enhances GAO’s ability to provide new and relevant reports and technical assistance to support Congressional informational needs 
Early-stage prototype
  • Large language model
  • Natural language processing
Performance Monitoring and ReportingQuantifies GAO's nonfinancial impacts and identifies unclaimed benefits from recommendations
  • Enhances reporting on GAO’s overall performance and accomplishments
  • Categorizes financial and nonfinancial benefits of agencies’ implementation of GAO recommendations
  • Matches GAO recommendations to resulting legislation
  • Enhances strategic planning through data-driven insights
Early-stage prototype 
  • Large language model
  • Natural language processing
  • Predictive modeling
Concept Induction ToolProduces meaningful concepts from unstructured text, such as open-ended survey responses or expert panel discussions
  • Increases efficiency and reduces manual qualitative data analysis
  • Creates interpretable and refinable concept groupings via human-in-the-loop iterative processes
Concept exploration
  • Large language model
  • Concept induction
IT AssistanceTriages IT help desk requests and answers internal GAO policy questions
  • Provides 24/7 self-service assistance to GAO employees
  • Allows support staff to focus on more complex requests
Concept exploration
  • Natural language processing
  • Large language model
  • Sentiment analysis
  • Integrated workflow and escalation

Source: GAO.

GAO Contacts