The Data Foundation recognizes the importance of the Senate's bipartisan efforts to develop a comprehensive framework for AI policy. However, it is crucial to emphasize the significance of data policy as a foundational element in the successful implementation of the roadmap's recommendations. As the Data Foundation has long advocated, the effectiveness of the proposed initiatives will largely depend on the availability of high-quality, accurate, reliable, and securely shared data.
Getting Data to "AI-Ready": Aligning the Roadmap with On-going Initiatives
The roadmap rightly encourages the authorization and funding of resources like the National AI Research Resource (NAIRR) through the passage of the CREATE AI Act (S. 2714). The effectiveness of the NAIRR in driving AI advancement depends on the availability of high-quality, diverse datasets that can be securely shared among researchers, government agencies, and other stakeholders. Importantly, the plan can go even further to include complementary initiatives, like the National Secure Data Service (NSDS), which is essential.
The NSDS aims to provide a secure framework for sharing sensitive and confidential data among government agencies, research institutions, and trusted, approved partners. The NSDS, already launched at the National Science Foundation and authorized by the CHIPS and Science Act of 2022, aims to develop a shared services model that streamlines and innovates data sharing and linking practices. If successfully implemented, this model will support and enable enhanced decision-making capabilities across all levels of government and various sectors.
The NSDS complements the Roadmap's emphasis on improving access to federal datasets to support AI innovation. By providing a secure framework for sharing sensitive data among government agencies and partners, the NSDS can play a critical role in operationalizing and enabling the safe and effective use of government data for AI purposes. This aligns with the Roadmap's call for an "all-of-government 'AI-ready data' initiative" and its recognition of the need to consider federal policy issues related to datasets used in AI development, including privacy concerns and potential transparency requirements.
The NSDS's focus on streamlining data sharing and linking across government levels and sectors resonates with the roadmap's emphasis on collaboration and breaking down silos. As the Roadmap highlights the importance of interdisciplinary collaborations that harness AI across academia, industry, national laboratories, and federal agencies, initiatives like the NSDS can facilitate the secure sharing of data necessary for such collaborations to thrive.
Several mechanisms are already in place to help the government develop a healthy data ecosystem that can support AI. The Foundations for Evidence-Based Policymaking Act (Evidence Act), which reauthorized the Confidential Information Protection and Statistical Efficiency Act and created a presumption of accessibility, provides a secure framework for sharing sensitive data among government agencies and partners. This framework supports initiatives like the NSDS, which can play a critical role in operationalizing and enabling the safe and effective use of government data for AI purposes.
Policy Recommendations for a Cohesive Data Governance and AI Framework
Chief Data Officers (CDOs) play a crucial role in supporting "AI-ready data" by ensuring data quality, accessibility, and usefulness for AI development. CDOs establish sound data governance and management practices, facilitate collaboration and knowledge sharing across departments, and develop enterprise-wide data strategies that support AI adoption while addressing ethical and regulatory concerns.
Recognizing the pivotal role of data policy in realizing the full potential of AI, the Data Foundation urges policymakers to take the following steps to bolster cohesive data governance across the federal government and facilitate responsible AI adoption:
-
Prioritize resources for complementary initiatives like the NAIRR and NSDS to ensure the availability of high-quality, diverse datasets for AI research and development.
-
Encourage collaboration between the NAIRR and NSDS to create a comprehensive framework for secure data sharing and AI innovation.
-
Support the work of CDOs across government agencies by providing clear guidance, resources, and authority to establish data governance practices that facilitate responsible AI adoption.
-
Promote the use and implementation of existing data-sharing frameworks, such as those established by the Evidence Act, to support the development of an "AI-ready data" ecosystem across the government.
Understanding Impact: Gathering Data and Evidence to Inform AI Policy
There are still pressing and urgent needs for data beyond feeding and developing AI itself. As AI rapidly advances and is adopted across industries, sectors, and governments, new policy questions will continually emerge that require rigorous data and evidence to address. Policymakers, including Congress, should proactively encourage ongoing information gathering to stay apprised of AI's evolving impacts and guide policymaking.
The Roadmap highlights several promising examples of this approach, such as the proposed Workforce Data for Analyzing and Tracking Automation Act (S. 2138), and calls for the Energy Information Administration to gather data on the energy usage of data centers and supercomputing clusters that power AI. Congress should identify additional opportunities to leverage the government's data gathering capabilities to inform AI policymaking, supporting projects at statistical agencies, voluntary data sharing partnerships with the private sector, and academic partnerships.
By continuously expanding the base of knowledge on AI's real-world impacts, Congress can proactively identify policy issues as they arise and develop evidence-based solutions that ensure the technology benefits the American people.
Moving Forward with a Solid Foundation
The Senate AI Policy Roadmap provides a solid foundation for addressing the challenges and opportunities presented by AI. However, the success of the Roadmap's recommendations will largely depend on the availability of high-quality, diverse, and securely shared data. The Data Foundation strongly urges policymakers to prioritize data policy as a critical component of the overall AI policy framework.
By investing in initiatives like the NAIRR and NSDS, supporting the work of CDOs, and promoting the use of existing data-sharing frameworks, Congress can ensure that the necessary data infrastructure is in place to support responsible AI innovation. By actively gathering evidence on AI's impacts through surveys, partnerships, and academic collaborations, policymakers can proactively address emerging policy issues and develop evidence-based solutions.
As the world moves forward in the era of AI, it is essential to recognize that data policy and AI policy are inextricably linked. By prioritizing data policy alongside the recommendations outlined in the Senate AI Policy Roadmap, we can create a solid foundation for responsible, innovative, and impactful AI development that benefits all.