KKMatch studies how people communicate, decide, and relate under real conditions — and publishes what the data actually shows. We do not publish findings we do not have.
Conversational behaviour: how people respond to uncertainty, signal disagreement, disclose trust, expand on a topic, or disengage. We study these as interaction patterns, never as diagnoses.
Observation first, inference second. Every published claim must be traceable to an observed signal, carry a confidence level, and state at least one alternative explanation. Raw media is not retained by default; features are.
We do not publish conclusions about medical conditions, mental health, race, religion, political belief, sexual orientation, or socioeconomic status. We do not publish findings we cannot evidence.
We are in the data-collection phase. This page will host our first genuine write-ups — for example what we learn from the first hundred completed Human Mirror sessions — as soon as that data exists and has been validated. Until then we would rather say nothing than guess.
Q: Can researchers access the dataset?
A: Not yet. Any future access will require explicit participant consent for research use and an ethics review.
Q: Is my conversation included?
A: Only if you explicitly grant research consent. It is off by default and can be withdrawn at any time.
Q: Do you publish raw video?
A: No. Raw media is not stored by default, and is never published.