Institute for AI Policy and Strategy (IAPS)

Shaping AI’s impact on security, innovation & global affairs.

[Institute for AI Policy and Strategy (IAPS) @iapsAI]

6/6 With thanks to authors @CassiaKing1 and @onni_aarne for their work on the report, which can be read here: https://t.co/KeTSglSaa2

[Institute for AI Policy and Strategy (IAPS) @iapsAI]

1/6 Maintaining a lead in the development of frontier AI models is essential if the U.S and its allies are to harden critical infrastructure networks and defend against national security threats from adversaries.

[Institute for AI Policy and Strategy (IAPS) @iapsAI]

5/6 As AI data center construction accelerates across Asia, Washington needs a strategy to address risks posed by remote access. Our new report recommends that BIS publishes a cloud compute rule. It also provides a policy escalation framework to address the issue.

[Institute for AI Policy and Strategy (IAPS) @iapsAI]

7/7 Thanks to authors @jan__wehner and @JKraprayoon for their work on our new report, Honeypots for AI Agent Insider Threats, which can be read here: https://t.co/BeQncVwiaC

[Institute for AI Policy and Strategy (IAPS) @iapsAI]

1/7 AI agent adoption at scale presents a new version of an old security problem—the insider threat. As enterprises and government agencies deploy AI agents to perform complex tasks, e.g., writing code, managing infrastructure, handling sensitive information, new risks emerge.

[Institute for AI Policy and Strategy (IAPS) @iapsAI]

6/7 To ensure this strategy keeps pace with advancing agent capabilities, AI assistance for designing and deploying honeypots is worth investing in. Our report outlines a technical research agenda and suggests production pilots to improve effectiveness and adoption of honeypots.

[Institute for AI Policy and Strategy (IAPS) @iapsAI]

4/4 Thanks to @theobearman for his work on the memo: https://t.co/hLLwmTPGCY

[Institute for AI Policy and Strategy (IAPS) @iapsAI]

1/4 Following a string of recent incidents involving rogue agent activity, both Anthropic and OpenAI have committed to embedding third-party evaluators—independent external experts who assess safety, security, and operational practices—within their companies.

[Institute for AI Policy and Strategy (IAPS) @iapsAI]

3/4 Our new memo outlines five foundations for voluntary arrangements to be built on. Companies cannot be left to “mark their own homework.” Under the right contract terms, embedded evaluators can provide a second line of defense, identifying serious risks before they cause harm.

[Institute for AI Policy and Strategy (IAPS) @iapsAI]

6/6 Find out how federal pilots could be used to test these systems in real-world conditions and develop operational practices. With thanks to @covinstantinop, @JKraprayoon, and @MattInThemittel for their work on the report: https://t.co/GPiBnR9uuu

[Institute for AI Policy and Strategy (IAPS) @iapsAI]

1/6 Offensive cyber agents, AI systems that directly and autonomously carry out partial or complete cyber attacks, are an emerging national security threat.

[Institute for AI Policy and Strategy (IAPS) @iapsAI]

5/6 This emerging AI application could multiply the effectiveness of human teams, drive down costs, enable more testing, help identify vulnerable systems and attack paths faster, and match the speed of agents.