International AI Safety Report 2025: First Key Update: Capabilities and Risk Implications

TL;DR

The 2025 AI Safety Report highlights significant advances in AI capabilities driven by new training techniques, improving performance across domains but raising new safety, reliability, and risk management challenges.

Contribution

This report provides the first major update on AI capabilities and associated risks since the initial publication, emphasizing recent technical improvements and their implications for safety.

Findings

AI systems now solve more complex problems across domains.

Performance on benchmarks in coding, math, and science has improved.

Reliability issues remain, with systems excelling in some tasks and failing in others.

Abstract

Since the publication of the first International AI Safety Report, AI capabilities have continued to improve across key domains. New training techniques that teach AI systems to reason step-by-step and inference-time enhancements have primarily driven these advances, rather than simply training larger models. As a result, general-purpose AI systems can solve more complex problems in a range of domains, from scientific research to software development. Their performance on benchmarks that measure performance in coding, mathematics, and answering expert-level science questions has continued to improve, though reliability challenges persist, with systems excelling on some tasks while failing completely on others. These capability improvements also have implications for multiple risks, including risks from biological weapons and cyber attacks. Finally, they pose new challenges for monitoring and controllability. This update examines how AI capabilities have improved since the first Report, then focuses on key risk areas where substantial new evidence warrants updated assessments.