Our Workstreams
01
Gold Standard Articulation
Articulating what frontier AI auditing should look like in the long-term (the “gold standard”) and why it’s important to achieve that gold standard.
Informs actions of policymakers, companies, philanthropists, and researchers.
02
Audit Research & Engineering
Conducting pilot assessments that push the envelope in terms of access, rigor, scope, and other dimensions. Producing open-source tools and training materials to enable reaching the gold standard more easily.
Raises the bar for quality while decreasing costs.
03
Policy & Advocacy
Analyzing and advocating for policies that incentivize rigorous auditing through mechanisms such as insurance, regulation, procurement criteria, and investor due diligence.
Builds demand by making frontier AI audits more economically appealing for various stakeholders.
Recent Work
RESEARCH
Frontier AI Auditing
A comprehensive vision for rigorous third-party verification of frontier AI developers’ safety and security claims and evaluation of their systems and practices against relevant standards, drawing on deep, secure access to non-public information.
RESEARCH
BenchRisk
We produced a metaevaluation tool, methodology, and results to assess the risk associated with relying on benchmarks for real world decisions of consequence. We named this tool suite "BenchRisk."
Collaborate with us
Interested in partnering on research or policy initiatives? Let's talk.