# The Foreign Policy AI Evaluation Gap

arXiv: https://arxiv.org/abs/2607.02955 · HTML: https://mlsanity.com/paper/2607.02955/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 19 of its 39 references that are in the map.
- First posted on arXiv 3 Jul 2026.

## Builds on (most cited first, 19 of 19)

- [AI Agents That Matter](https://mlsanity.com/paper/2407.01502/) (2024)
- [Improving Language Model Negotiation with Self-Play and In-Context Learning from AI Feedback](https://mlsanity.com/paper/2305.10142/) (2023)
- [Lessons from the Trenches on Reproducible Evaluation of Language Models](https://mlsanity.com/paper/2405.14782/) (2024)
- [Model evaluation for extreme risks](https://mlsanity.com/paper/2305.15324/) (2023)
- [Toward Trustworthy AI Development: Mechanisms for Supporting Verifiable Claims](https://mlsanity.com/paper/2004.07213/) (2020)
- [Sociotechnical Safety Evaluation of Generative AI Systems](https://mlsanity.com/paper/2310.11986/) (2023)
- [Outsider Oversight: Designing a Third Party Audit Ecosystem for AI Governance](https://mlsanity.com/paper/2206.04737/) (2022)
- [Escalation Risks from Language Models in Military and Diplomatic Decision-Making](https://mlsanity.com/paper/2401.03408/) (2024)
- [Rethinking Model Evaluation as Narrowing the Socio-Technical Gap](https://mlsanity.com/paper/2306.03100/) (2023)
- [Human vs. Machine: Behavioral Differences Between Expert Humans and Language Models in Wargame Simulations](https://mlsanity.com/paper/2403.03407/) (2024)
- [Emergent Bartering Behaviour in Multi-Agent Reinforcement Learning](https://mlsanity.com/paper/2205.06760/) (2022)
- [Reality Check: A New Evaluation Ecosystem Is Necessary to Understand AI's Real World Effects](https://mlsanity.com/paper/2505.18893/) (2025)
- [Benchmarking LLMs for Political Science: A United Nations Perspective](https://mlsanity.com/paper/2502.14122/) (2025)
- [Critical Foreign Policy Decisions (CFPD)-Benchmark: Measuring Diplomatic Preferences in Large Language Models](https://mlsanity.com/paper/2503.06263/) (2025)
- [Democratizing Diplomacy: A Harness for Evaluating Any Large Language Model on Full-Press Diplomacy](https://mlsanity.com/paper/2508.07485/) (2025)
- [A Deep Reinforcement Learning Approach to Concurrent Bilateral Negotiation](https://mlsanity.com/paper/2001.11785/) (2020)
- ["ChatGPT, Don't Tell Me What to Do": Designing AI for Context Analysis in Humanitarian Frontline Negotiations](https://mlsanity.com/paper/2410.09139/) (2024)
- [Advancing AI Negotiations: A Large-Scale Autonomous Negotiation Competition](https://mlsanity.com/paper/2503.06416/) (2025)
- [DipLLM: Fine-Tuning LLM for Strategic Decision-making in Diplomacy](https://mlsanity.com/paper/2506.09655/) (2025)

## Abstract (arXiv)

We argue that AI systems used in conducting foreign policy tasks - broadly enacting 'statecraft' - should be a priority test case for technical AI governance research. In enacting foreign policy, we refer to the formulation and implementation of external objectives by political actors. Statecraft is a high-consequence deployment domain, with extreme downside risks and structural properties that standard evaluation practices handle poorly. These features include partial observability, unbounded action spaces, contested ground truth, and multidimensional objectives. This paper advocates for a literature-grounded research agenda. Our contribution is threefold: (i) a claim about the structural conditions of foreign policy that combine catastrophic tail risk with technical evaluation complexities, (ii) an ECOSYSTEM review that highlights the asymmetric focus on ASSESSMENT features over ACCESS, VERIFICATION, SECURITY, and OPERATIONALIZATION, and (iii) a demand-side evaluation framework that decomposes foreign-policy workflows into bounded, evaluable sub-tasks with human recombination. As AI systems are already being deployed in the conduct of war and peace, amid limited public evaluation infrastructure from the technical AI governance community, this agenda is an urgent priority.