RESEARCH

Exploring Human Fluency, in People and Artificial Intelligence.

Can an AI form an accurate model of a specific person? And if it can — what is that accuracy worth? The Human Fluency Benchmark explores the first, under independent academic governance. The Fluent Interaction Conjecture tests the second, in the open, pre-registered, published whichever way the data lands. The AI Interpreter examined the accuracy of Frontier Models and their validity and reliability in interpreting multiple psychometric instruments. Each stands on its own evidence.