These are miscellaneous resources related to what labs should do and what they are doing.
They are mostly disconnected from the main part of this site.
Other scorecards collects other work evaluating labs on AI safety.
Commitments collects companies' commitments relevant to AI safety.
Companies' policy advocacy collects companies' policy recommendations and engagement with policymakers.
Open letters lists open letters related to AI safety.
Integrity incidents collects instances of low-integrity behavior from AI companies.
At AI Safety Claims Analysis, I write about companies' model evals for dangerous capabilities.