Regulatory Consulting at Institute for AI Policy and Strategy (IAPS)

Overview

The Institute for AI Policy and Strategy (IAPS) is a nonpartisan nonprofit focused squarely on reducing risks from frontier AI — from today’s most capable models toward the possibility of transformative future systems. It is small by design and unusually credible for a specific reason: it accepts no funding from for-profit organizations, protecting the independence of its work. IAPS is organized around three teams — Frontier Security, Compute Policy, and International Strategy & China — and it deliberately bridges the technical and policy worlds, with staff who understand both frontier models and the hardware underpinning them, distributed across Washington, San Francisco, London, and Oxford. The institute does both deep, agenda-driven research and rapid, decision-maker-facing briefings, and it produces public and non-public work depending on the sensitivity of the topic. Its Compute Policy team has become one of the most substantive voices on AI-hardware governance, exploring how existing U.S. export controls actually function and what technical mechanisms could enforce them — including on-chip governance and chip location verification. For anyone entering the field, IAPS is also one of the best on-ramps that exists: its fully funded AI Policy Fellowship places professionals alongside senior experts on real frontier-AI governance projects and explicitly aims to help them build the skills, domain knowledge, and connections for an AI-policy career in think tanks, government, or beyond.

AI-Focused Regulatory Work

IAPS is small, sharp, and unusually credible — and its niche is the most technical corner of AI policy, where chip engineering meets export-control law. Its flagship contribution, ‘Secure, Governable Chips’ (produced with the Center for a New American Security), introduced the concept of ‘on-chip governance’: the idea that security features built directly into AI chips could help mitigate national-security risks from broadly capable dual-use systems while still protecting user privacy. From there its Compute Policy team — including researchers like Erich Grunewald and Onni Aarne, and led on compute by an associate director who was formerly Head of AI Policy at His Majesty’s Treasury — has built out a body of work on location-verification mechanisms for AI chips, evidence that advanced chips are being smuggled into China, the security of AI data centers, and close reading of specific policy moves like the December 2025 decision to permit H200 exports to China in exchange for a licensing fee. What makes IAPS distinctive is the combination of technical seriousness and institutional independence: because it refuses for-profit funding and collaborates with organizations like CNAS, GovAI, and The Future Society, its recommendations carry weight precisely because they can’t be dismissed as industry-shaped. And unusually for a research shop, IAPS is explicitly a talent-building institution — its fellowship has trained fellows who have gone on to work at Oxford’s AI Governance Initiative, The Future Society, and beyond, making it one of the highest-leverage entry points into a frontier-AI-policy career in the entire directory.

Notable AI Policy Work

• Co-published ‘Secure, Governable Chips’ (with CNAS), introducing ‘on-chip governance’ — security features on AI chips to mitigate national-security risks from dual-use systems while preserving privacy.
• Built a leading Compute Policy research program on technical enforcement of U.S. export controls, including chip location-verification mechanisms (privacy-preserving and tamper-resistant designs).
• Documented evidence that substantial quantities of advanced AI chips are being smuggled into China, and analyzed AI data-center security against sophisticated adversaries.
• Analyzed specific export-control policy (e.g., the December 2025 decision to permit H200 sales to China in exchange for a 25% export fee), translating fast-moving policy into decision-maker guidance.
• Runs three research teams — Frontier Security, Compute Policy, and International Strategy & China — producing both deep research and rapid briefings, public and non-public.
• Operates a fully funded AI Policy Fellowship placing professionals with senior experts and building a talent pipeline into AI-policy careers (fellows have gone on to Oxford’s AI Governance Initiative, The Future Society, and more).

Thought Leadership

1. IAPS Compute Policy Hub
   IAPS’s Compute Policy Hub — research on technical enforcement of export controls, chip location verification, and on-chip governance.
2. Compute Governance research
   IAPS’s compute-governance research collection, including ‘Secure, Governable Chips’ (with CNAS) and chip-smuggling analysis.
3. 2025 AI Policy Fellows
   The IAPS AI Policy Fellowship cohort — a fully funded pipeline into frontier-AI-policy careers in think tanks and government.
4. About IAPS (mission & independence policy)
   IAPS’s mission and independence policy — nonpartisan, frontier-AI-focused, and funded without for-profit money.

Clients

Serves frontier-AI and national-security policymakers in the U.S., U.K., and internationally, including on export controls and compute governance. Nonpartisan and independent — accepts no for-profit funding — with frequent collaborators including CNAS, GovAI (Oxford), and The Future Society.