Learning path · For journalists, researchers, students, and policy readers
Checking claims about AI
Read benchmark results, safety documents, content labels, data disclosures, infrastructure announcements, and policy texts with care.
6 steps · about 32 min of reading
- Step 1 · Explainer · Intermediate · 5 minHow to read an AI evaluationWhat a benchmark result can and cannot show.
- Step 2 · Explainer · Intermediate · 6 minHow AI developers test models for safetySystem cards, red teaming, and published safety frameworks.
- Step 3 · Explainer · Beginner · 6 minLabels, watermarks, and content credentialsLabels, watermarks, and content credentials.
- Step 4 · Explainer · Intermediate · 5 minUnderstanding training data disclosuresWhat a provider reveals about training data.
- Step 5 · Explainer · Intermediate · 5 minReading AI infrastructure and energy claimsPlans, power, and operating capacity.
- Step 6 · Explainer · Intermediate · 5 minFinding AI policy and standards sourcesFinding and dating the original policy text.
Support Us
Help keep USASI useful.
Optional. No USASI account required. Payment takes place on the linked provider’s website (Buy Me a Coffee).
About supporting this project