This analysis tracks global large language model (LLM) adoption across 135 countries, revealing how pricing, multilingual capabilities, and AI diplomacy shape the evolving U.S.–China competition. The findings provide critical insights for policymakers and industry leaders navigating the geopolitical and economic implications of AI market dominance.
How can the U.S. government and frontier artificial intelligence model developers strengthen the industry's security practices? We identify four practicable governance approaches to help navigate these challenges.
This report assesses eight frontier large language model (LLM) agents on their ability to design DNA segments, interact with a benchtop DNA synthesizer, and generate laboratory protocols—tasks with potential dual-use implications. While performance varied, all tested models successfully designed biologically coherent DNA segments in some attempts, highlighting both technical progress and biosecurity concerns.
This report introduces a dual-component risk-scoring tool to assess the potential misuse of AI in modifying viral capabilities such as host range, replication, and immune evasion. By evaluating both the biological impact of modifications and the technical skill required to achieve them, the tool provides a foundation for developing redlines and decision frameworks that balance innovation with biosecurity.
CAST is now recruiting for a Senior Research Lead - AI Security Portfolio, who will apply strong management acumen and expertise in AI security to shape strategy and lead high-performing teams. If you or someone you know would be a good fit, we encourage you to apply.
Thank you for following our work. We look forward to sharing more updates and insights with you in the months ahead.