Center for Security and Emerging Technology

Center for Security and Emerging Technology (CSET) is a policy research organization within Georgetown University's Walsh School of Foreign Service. It studies the security implications of AI, advanced computing and biotechnology. Its work combines technical and foreign-language research, data analysis and engagement with decision-makers. This profile covers the center's AI-policy contributions rather than Georgetown's activities as a whole.

Institutional home: Georgetown University’s Walsh School of Foreign Service.

Research focus: AI, advanced computing and biotechnology.

Funding model: Philanthropic support, with a published donor list and independence policy.

Official site: cset.georgetown.edu

Last source review: September 12, 2026

Research and policy role

CSET's research encompasses AI applications, assessment, cybersecurity, computing supply chains, competition and workforce questions. It publishes reports, testimony, translations and data products. Its Emerging Technology Observatory makes research resources available to outside users, while its Government AI HUB examines federal understanding and use of AI. The center's approach gives policymakers evidence about the institutions and resources underpinning AI as well as about models themselves.

CSET's funding policy distinguishes institutional analysis from lobbying. It says research priorities respond both to policymakers' questions and researchers' identification of gaps. Although it describes itself as not advocating particular policies, it expressly allows publications to include researchers' assessments of actions that would promote security. Recommendations should therefore be attributed to the relevant report or submission rather than assumed to be a uniform position on every related controversy.

Critical infrastructure

The October 2024 report Securing Critical Infrastructure in the Age of AI synthesizes a June workshop. It distinguishes vulnerabilities introduced when operators deploy AI from malicious uses of AI against infrastructure. It identifies unequal resources among operators as a governance problem: a small provider may lack staff and evaluation capacity available to a large enterprise.

The report recommends assistance and information sharing, clearer responsibility for AI risk inside organizations and better integration of AI and cybersecurity guidance. It also calls for clarity about incident definitions and reporting channels. Its discussion of the 2023 executive order is historical context for the workshop, not a statement that all of that order's provisions remain operative. The recommendations address practical capacity as well as formal compliance.

AI Action Plan submission

CSET's March 2025 recommendations address US research leadership, competition with China and the benefits and risks of adoption. The submission supports sustained public research funding, interdisciplinary work and widely usable datasets. It also advocates more competitive cloud markets, scrutiny of restrictive contracting and review of mergers and corporate partnerships.

For open models, the submission recommends avoiding restrictions unless clear, measurable risk thresholds are exceeded. It supports resources for examining both benefits and risks, rather than treating openness as either inherently harmless or inherently unacceptable. The submission is a documented intervention in the Action Plan process; the presence of similar language in a later government document would not alone prove that CSET caused its adoption.

International AI infrastructure

Pablo Chavez's October 2025 U.S. AI Statecraft examines four US-linked AI infrastructure partnerships in the United Arab Emirates and Saudi Arabia. It argues that individual commercial transactions need coherent governance, oversight and alignment with strategic objectives. The report treats technology infrastructure as a potential channel for shaping international rules and practices.

The analysis relies on public documents and reporting and identifies uncertainty about some underlying agreements. It distinguishes announced partnerships from enforceable oversight arrangements. Its recommendations are the author's analysis of how the United States could structure such relationships, rather than evidence that every proposed safeguard was incorporated in an executed agreement.

Funding and affiliations

CSET publishes donor categories that include foundations, individuals, companies and other sources. Open Philanthropy appears in its highest displayed contribution band; other named supporters include Google.org, the Hewlett Foundation and the Frontier Model Forum AI Safety Fund. The reviewed page does not provide a single reporting period for all displayed bands, so they should not be read as annual totals. CSET also describes conflict-of-interest disclosure, review and mitigation under Georgetown policies.

Related articles

• Policy
• United States policy on catastrophic AI risk
• Centre for the Governance of AI
• Stanford Institute for Human-Centered Artificial Intelligence