AI Policy Tools for Today: A Bold Investment in NIST

At the core of AI policy lies the ability to accurately describe and quantify the capabilities and risks of systems. To this end, Anthropic proposes a substantial increase in federal funding for NIST (National Institute of Standards and Technology) to support AI measurement and standardization work.

As AI technology advances rapidly, there are increasing cases where unexpected risks arise during model training or are discovered only after deployment. Rather than complex alternatives such as establishing new regulatory bodies or pausing development, strengthening the capabilities of NIST, which already possesses proven technical standards, could be a more practical and immediate solution.

NIST has already played an important role in the field of AI measurement through initiatives such as the Face Recognition Vendor Test (FRVT) and the AI Risk Management Framework (AI RMF). However, budget support in this area has stagnated relative to the recent pace of technological advancement.

With bold investment, NIST could build foundational measurement technologies and standardize them across the industry. It could also build community resources such as testbeds to assess AI capabilities and risks, thereby supporting safe innovation.