Navigating Federal Data Policy: Perspectives for Congressional Staff

The following is an overview of a section of a class lecture I offered to congressional staff in August 2025 as part of the USA Facts, University of California Berkeley Data Skills for Congress course.

Data in government is like plumbing—you always need it to work, but when it breaks, everyone notices. As artificial intelligence dominates headlines and debates rage about data's role in society, it's worth stepping back to understand the complex framework of laws and regulations that govern how the federal government collects, manages, and shares information.

Evidence Informs, But Doesn't Dictate

The foundation of data policy rests on a crucial principle: evidence is an objective description of what is, while policymaking is the subjective determination of what ought to be. This distinction matters because while data and analysis should inform decisions, they never dictate them. We want our data and evidence to have a seat at the table for decision-making. Values, elections, and democratic processes remain central to governance.

Consider unemployment rates. Is 4.5% too high or too low? The answer depends entirely on your perspective and priorities. Data provides context, but human judgment—guided by values and electoral mandates—makes the decision.

A Scattered Legal Landscape

Federal data policy isn't found in one place. Instead, it's dispersed across dozens of laws spanning more than five decades. The Privacy Act of 1974 still forms the backbone of U.S. privacy law. The DATA Act of 2014 standardized spending information across agencies. The Foundations for Evidence-Based Policymaking Act of 2018 (Evidence Act) created Chief Data Officers and established a default that government data should be open. There are many more laws and policies beyond this. 

This fragmentation creates complexity. The United States now has more than 3,000 privacy laws when including state and local jurisdictions. Add emerging AI regulations, and the landscape becomes even more intricate. Yet this complexity also creates opportunities for congressional staff who can help connect these authorities and find synergies across different policy areas.

Success Stories and Implementation Challenges

The DATA Act demonstrates how good data policy works in practice. By requiring standardized spending information across all federal agencies, it enabled the creation of USAspending.gov—making the same financial data available to both congressional staff and citizens in rural Missouri. When the Department of Government Efficiency (DOGE) made federal spending more transparent recently, they relied on infrastructure created by this bipartisan 2014 law.

However, implementation remains uneven. The Government Accountability Office has found that while most agencies have vastly improved their spending information quality, the Department of Defense still has significant room for improvement.

The Evidence Act: A Comprehensive Framework

The most comprehensive recent data policy legislation emerged from the U.S. Commission on Evidence-Based Policymaking, created through bipartisan leadership from then House Speaker Paul Ryan and Sen. Patty Murray. The Evidence Commission recognized that Congress often lacked the information needed for major program decisions.

The resulting Evidence Act focused on three key themes:

• Capacity: New evaluation officers across major agencies develop learning agendas that identify priority research questions. These agendas represent agencies acknowledging, with some humility, what they need to know for upcoming regulatory decisions, reauthorizations, and major policy actions. Developing and using evidence requires people and resources for effective decision-making—the capacity investments were designed to address this fundamental need.
   
• Governance: Chief Data Officers at every agency focus on improving data quality and accessibility. These positions didn't exist in government a decade ago but are now well-established, with many also taking on AI oversight responsibilities. Data governance relates directly to making information both high quality and machine-readable. When these elements work together, they create machine-understandable data for the AI era.
   
• Privacy: One of the world's strongest privacy laws—the Confidential Information Protection and Statistical Efficiency Act (CIPSEA)—sets both civil and criminal penalties for violating pledges of confidentiality to the American people. CIPSEA serves as the foundation for the U.S. federal statistical system, which forms the backbone of much of our country's core data collection and analytical activities.

Current Challenges and Opportunities

Recent months have brought significant changes to federal data capacity, reflected in the Data Foundation’s Evidence Capacity Pulse Reports. Approximately 40-50% of the evaluation officers appointed under the Evidence Act have retired or left their positions. Additionally, there has been 10-20% attrition across the federal statistical system. Substantial turnover among chief data officers. This represents both a challenge and an opportunity to rebuild and transition with both fresh perspectives and approaches.

The Open Government Data Act, passed as part of the Evidence Act, finally received implementation guidance in January 2025—just five days before the Biden administration ended. This guidance remains in effect, providing a framework for agencies to make more data publicly available while protecting confidential information. The first regulation required by CIPSEA was finalized in late-2024, with more included in the current regulatory agenda to bolster the statistical infrastructure. Open data plans are being published from chief data officers. Revisions to privacy policy and law are underway, including reimagining expanded applications of routine uses of administrative data. 

The rapidly evolving changes in the system must also be paired with clarity of law and policy – alongside clear and compelling transparency about the changes and how the American people’s data are being used. 

Looking Forward: AI and Data Policy Convergence

The AI revolution creates both urgency and opportunity for better data policy. Much of what we call "AI policy" is fundamentally data policy—algorithms are only as good as the data that trains them. The government possesses rich information that could enable AI innovation, but only if that data is machine-readable and machine-understandable.

The Evidence Act's framework, if fully implemented, would create the foundation for this AI-enabled future. But we're years behind where we should be, requiring renewed investment, attention, and leadership in 2025 and beyond.

Practical Next Steps for Congressional Staff

For congressional staff working on oversight or policy development, several concrete opportunities exist:

• Learning Agendas: Federal agencies are publishing research priorities that could benefit from academic collaboration or congressional attention. Congressional staff can offer perspectives about questions and priorities to include in agency learning agendas.
• Chief Data Officers: These relatively new positions offer direct access to agency data expertise and can help identify high-priority open data assets. Congressional staff can meet with CDOs and ask questions about data governance in meetings or hearings. Statistical agencies should also be part of these discussions, including to build greater understanding about the role of surveys, administrative data, and alternative data for answering relevant questions.
• Grant Outcomes: The administration is developing a system for improving on analyzing and reporting grant outcomes, building on decade-old legislation around transparent grant reporting. Staff with grants in their portfolio of programs can provide resources and direction to programs about congressional priorities for key outcomes of interest.
   

The federal data ecosystem has strong foundations, established through bipartisan legislation over many years. The challenge now is ensuring these frameworks are fully implemented and adapted for an AI-enabled future. Congressional staff have both the opportunity and responsibility to ensure these critical information systems serve democratic governance effectively.

Data policy may not typically grab headlines like other technology issues, but it underpins virtually every aspect of government operations and public accountability. Getting it right matters for both effective governance and public trust.

NICK HART, President and CEO of the Data Foundation.