The report argues that the rapid integration of AI creates novel and complex risks—such as deepfakes, misinformation, and intellectual property disputes—that do not fit within existing insurance lines, challenging the technology's overall insurability. Evidence shows that while lawsuits center on IP, the scope of harm is expanding into privacy and product liability, and insurers are responding unevenly with a mix of exclusions and limited coverage. Policymakers must therefore focus on managing systemic, correlated AI risks, including shared model dependencies and universal attacks, to prevent market failure and guide necessary regulatory oversight.
Breaking the deadlock on AI governance
English Summary
Chatham House argues that progress on international AI governance is currently stalled due to geopolitical tensions, institutional weaknesses, and public-private imbalances, suggesting that a significant AI-related crisis may be the only catalyst for rapid, binding global cooperation. The paper draws lessons from past crises like the 2008 financial crisis and the WannaCry attack, highlighting the importance of pre-existing institutions and technical expertise for effective crisis-driven governance. Consequently, policymakers should prioritize investing in foundational AI governance infrastructure now to be prepared for a potential crisis and facilitate a robust response.
Related Entries
-
1.
-
2.
The article argues that China is establishing digital chokepoints by penetrating critical U.S. infrastructure, creating a cyber asymmetry that gives Beijing a strategic advantage. This vulnerability is compounded by the decentralized nature and outdated hardware of American utilities. To counter this, the U.S. must leverage AI to create 'digital twins'—virtual replicas of critical systems—which allow for proactive stress-testing and vulnerability mapping. Policy recommendations include establishing a secure, national enclaved for AI testing, coupled with legal reforms (like safe harbor) and government-private partnerships to accelerate the necessary upgrades and secure the digital frontier.
-
3.Open-Weight AI Models May Increase Biological Misuse Risks: Assessing Anti-Refusal Tampering, Capability Enhancement, and Publicly Available Uncensored Models (RAND)
The RAND report argues that open-weight Large Language Models (LLMs) significantly increase biological misuse risks because their public availability allows for unauthorized modification and removal of safety safeguards. Key evidence shows that moderately skilled actors can cheaply and reliably tamper with these models—primarily by removing refusal mechanisms—thereby making existing dual-use biological knowledge freely accessible without developer controls. This presents a critical policy challenge: balancing the substantial benefits of open research and transparency against the risk of unrestricted redistribution in sensitive domains like biosecurity. Policymakers must therefore develop strategies to mitigate the threat of tampering and misuse while preserving the scientific and economic advantages of open-source AI.
-
4.
The report argues that public procurement is a useful, yet bounded, governance instrument for middle powers dependent on foreign frontier AI. By embedding contractual guarantees—such as model access for evaluation and termination clauses—into state contracts, governments can mitigate risks like vendor lock-in and data sovereignty vulnerabilities. However, the analysis finds that individual state demand provides limited leverage; therefore, the key strategic recommendation is that middle powers must coordinate their procurement efforts. This collective action is necessary to aggregate demand, share evaluation capabilities, and strengthen their collective bargaining position against dominant AI providers.
-
5.
The rapid financing of the AI boom through massive corporate debt issuance is creating significant stress on global financial markets. This influx of private capital forces competition with government Treasury bonds, as AI companies offer higher yields than equivalent sovereign debt, thereby pushing up long-term interest rates. While this signals strong investment demand for AI, it raises concerns about systemic risk and the potential destabilization of core bond markets. Policymakers must navigate the tension between fueling critical technological growth and maintaining stable public borrowing costs to prevent a financial crisis.