A recent Science editorial argues we're losing the ability to understand AI. One piece of the problem is more tractable than it looks, if we stop measuring the model and start measuring the person.
The human-AI interaction is psychological in nature, and language is the vehicle that carries that information. AI systems already infer a user's cognitive and emotional state from language, but that inference lives inside the model, unvalidated and unverifiable. Receptiviti Labs measures it directly, using psycholinguistic methods validated across three decades of research. We built Receptiviti Labs to make AI systems safer and more aligned with the humans they've been devel
Receptiviti Labs
5 days ago
The measurement layer for the human side of AI.
API access · On-prem deployment · Research partnerships