Roundtable: Vertex Leasing, Usage, Purchase () Government AI Roundtable Series Summary Report

Key Chapter Title List

1. Introduction
2. Discussion Overview
3. Key Takeaways and Recommendations
4. Substantially Similar, Differing in Terminology
5. Investing to Save
6. Maximizing AI Benefits Through Transparency
7. Automation and Human-Machine Teams
8. Conclusion
9. Authors
10. Acknowledgments
11. Endnotes

Document Introduction

This report is a summary of the outcomes from the fourth and culminating session of the "Government AI: Hire, Use, Buy (HUB)" roundtable series, co-hosted by Georgetown University's Center for Security and Emerging Technology, the Beeck Center for Social Impact and Innovation, and the Institute for Technology Law and Policy. This series of roundtables, held in 2024, aimed to convene leaders from government, industry, civil society, and academia to delve into the legal liabilities, governance frameworks, and strategic challenges faced by the U.S. government in using, procuring AI tools, and attracting AI talent.

The report first reviews recent U.S. government policy actions in the field of artificial intelligence, including the Executive Order on the Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence signed by President Biden in 2023, and the government-wide policy issued by the Office of Management and Budget (OMB) in 2024. These efforts build upon earlier executive orders (such as Executive Order 13960 from the Trump administration), yet many critical issues regarding AI application, procurement, and talent recruitment remain inadequately addressed. Against this backdrop, this culminating session aimed to synthesize findings from the previous three roundtables and distill potential policy recommendations and a set of key challenges.

The core content of the report revolves around several key themes discussed during the meeting. First, participants analyzed the potential continuity of AI policy in the context of possible leadership changes within the U.S. government. The discussion noted that, despite potentially different phrasing, there is significant commonality in the efforts of both administrations across many substantive AI policy areas, such as agency AI use inventory work—"substantially similar, differing in terminology." This suggests that certain foundational work may persist beyond political cycles.

Second, the report emphasizes the upfront resource investment required for effective AI integration. Participants generally agreed that, in the face of potential government budget cuts, sufficient investment in time, funding, and personnel—such as conducting comprehensive and effective AI use inventorying and dataset organization—can save resources and enhance efficiency in the long run. Regarding talent recruitment and retention, the report points to the need for high-level commitment, dedicated personnel, and providing valuable onboarding experiences for new hires to overcome bureaucratic hurdles in government hiring.

Transparency was established as a core principle for building public trust, thereby securing the social and political capital necessary for the government's large-scale application of AI. The report criticizes current fragmented and formalistic disclosure practices and advocates for establishing a government-wide, integrated, and centralized AI information disclosure mechanism to enable better public and civil society oversight and accountability.

Regarding the role of AI in government, the meeting explored whether it serves as an automation tool to replace human labor or as an enabler to enhance the work of existing officials. The discussion concluded that, while there is a strong motivation to pursue technological solutions in the name of efficiency, current AI technology is more likely to be used to empower government personnel, forming effective human-machine teams. Core government functions (such as contracting officers) still require human leadership, while automation replacement may occur more at the level of private contractors undertaking government outsourced work.

The report concludes by summarizing two main recommendations: first, to provide adequate time, funding, and personnel resources for AI integration work to promote long-term efficiency; second, to centralize transparency and disclosure reporting to better meet the public's right to know. The report acknowledges that numerous challenges remain in government AI integration and calls for continued dialogue and cooperation among the government, Congress, civil society, industry, and academia to ensure that the development of AI ultimately serves the public interest.