Institute for AI Policy and Strategy (IAPS)

There are open questions—how auditors operate in non-cooperative jurisdictions, how to incentivize quality over cost-cutting, and whether detection alone deters one-time smuggling schemes. We discuss these and more in the working paper. We're publishing this to get feedback. Read the full working paper: iaps.ai/research/expor…

We think this could work as a license exception: importers hire an auditor, and the export proceeds without a case-by-case BIS review. This keeps trade flowing while adding meaningful enforcement.

New working paper from @iapsAI: We propose "export auditors"—independent, for-profit firms accredited by BIS to help enforce U.S. export controls on AI chips and semiconductor manufacturing equipment. Here's the problem they'd solve, and how they'd work. 🧵

6/ Full memo here: iaps.ai/research/ai-di…

5/ Theo proposes three recommendations to support industry efforts to counter distillation attacks: (1) consider BIS Entity List designations for adversary AI companies conducting distillation attacks; (2) assess the merits of PAIP Act sanctions against those engaging in or facilitating distillation attack; (3) explore the development of a NIST-led AI Distillation Defense Framework for the broader ecosystem.

1/ New IAPS policy memo from @theobearman. AI Distillation Attacks: The Case for Targeted Government Intervention. Theo examines how distillation attacks work, why there is a case for targeted government intervention and what that might look like. In February 2026, Anthropic, OpenAI, and Google each published evidence of systematic campaigns by Chinese AI companies to extract capabilities from American frontier models.

Read the report: iaps.ai/research/highl…

🚨 What would it look like if AI systems could autonomously run cyber campaigns for weeks or months at the level of top-tier criminal groups & intelligence agencies? Our new report, "Highly Autonomous Cyber-Capable Agents," tries to answer this question and argues these systems could be technically feasible far sooner than most defenders are prepared for. Read the report: iaps.ai/research/highl…

Read the full blog post: iaps.ai/research/takea…

New IAPS blog post: Takeaways from the India AI Impact Summit India recently held the first AI Summit in the Global South. Attended by key AI executives and delegates from over 100 countries, the Summit provided insights into the US's and China's AI policy priorities, offered contrasting visions of AI sovereignty, and reinforced the continued shift away from frontier AI topics. IAPS analyzes the key outcomes and geopolitical dynamics at the Summit, as well as what to watch for in the coming months.