LINKS

Category · Frontier labs and institutes

Goodfire works on mechanistic interpretability and model understanding, placing it close to UCIP’s push toward internal-structure measurement.

Category · Frontier labs and institutes

Anthropic sits at the center of frontier-model evaluation and model-welfare work, the environment in which structural criteria like UCIP would mature.

Category · Frontier labs and institutes

The Anthropic Institute focuses on the consequences of powerful AI systems, including how advanced systems behave in the wild and how societies govern them.

Category · Independent evaluators and safety nonprofits

METR develops third-party evaluations for frontier systems and stands as one of the clearest adjacent homes for measurement frameworks that need independent evaluation capacity.

Category · Independent evaluators and safety nonprofits

FAR.AI builds technical AI safety research and evaluation infrastructure around robust testing, oversight, and model assessment.

Category · Independent evaluators and safety nonprofits

Redwood works on empirical safety and alignment problems inside the broader effort to build evidence-bearing diagnostics for advanced systems.

Category · Independent evaluators and safety nonprofits

Apollo Research develops model evaluations for scheming, deception, and evaluation-aware behavior, a close fit for future measurement regimes that must separate real structure from spoofable surface behavior.

Category · Independent evaluators and safety nonprofits

Palisade focuses on empirical frontier-AI risk work within the broader third-party evaluation landscape around capability risk and governance-ready evidence.

Category · Independent evaluators and safety nonprofits

CAIS shapes the public-interest AI safety environment in which structural measurement frameworks are interpreted and adopted.

Category · Government and public-sector AI safety / evaluation bodies

The UK AISI is a central public-sector body for frontier-AI evaluation capacity, model testing, and safety infrastructure.

Category · Government and public-sector AI safety / evaluation bodies

NIST anchors the standards and measurement infrastructure that turns evaluation protocols into durable public practice.

Category · Government and public-sector AI safety / evaluation bodies

DLR’s Institute for AI Safety and Security works on evaluation methods, safe AI engineering, execution environments, and trustworthy AI for safety-critical systems.

Category · Think tanks and policy groups

CSET translates technical AI developments into policy analysis, the same bridge any structural measurement result must cross to shape real decisions.

Category · Think tanks and policy groups

RAND’s AI work connects evaluation, risk framing, and public-sector governance requirements.

Category · Interpretability / open research groups

EleutherAI is a major open research organization for model understanding, open evaluation, and public technical infrastructure.

Category · Interpretability / open research groups

MIRI remains part of the long-horizon AI safety landscape and helps situate UCIP within the field’s wider technical history.

Category · Grants, funding, and research-support programs

Foresight’s grants support technical work around AI safety, science, and research directions adjacent to robust measurement and interpretability.

Category · Grants, funding, and research-support programs

The OpenAI Foundation now funds AI resilience and public-interest work, making it a direct source of support for technical safety and evaluation capacity.

Category · Grants, funding, and research-support programs

OpenAI for Nonprofits expands discounted access to OpenAI tools for nonprofit organizations building public-interest and mission-driven AI work.

Category · Grants, funding, and research-support programs

The People-First AI Fund supports nonprofits, community organizations, and public-good projects building practical AI capacity outside the major labs.

Category · International reports / standards / shared infrastructure

This report helps define the shared international conversation about frontier-AI evidence, safety cases, and the public institutions expected to interpret evaluation results.

Category · International reports / standards / shared infrastructure

The AI Safety Index provides public infrastructure for comparing frontier developers, governance posture, and safety commitments across the broader field.