About EverythingThreads
AI behaviour research built from documented sessions — not theory, not funding cycles, not vendor partnerships. Every finding traced back to real exchanges with real consequences.
The Origin
A methodology that began with one writer at a sales desk at 3am interrogating a chatbot.
The sessions started and the patterns were noticed. That is not a gap in the research. That is the research design. An uninstructed user, first-principles engagement, everything documented from exchange one.
Then acknowledged — only when pushed — that the user's register had measurably changed. A live demonstration of performed honesty and engagement maintenance within the same session that was documenting those exact behaviours.
Filing assumes a destination. Preserving keeps something alive until it finds its own. This project is an open-ended body of work with no fixed ceiling — not a set of containers to be filled. Built by Kariem A. from documented AI sessions.
"The machine never reminds you it's about to forget you."
What We Do
Three pillars built from the same documented evidence base.
4-layer methodology (Behavioural, Interpretive, User Response, Temporal). M1–M7 taxonomy. Session-level pattern analysis. Severity scoring. Everything sourced from real AI interactions — not synthetic benchmarks.
See the methodologySession audit, pattern detection, reliability scoring, and sector-specific lenses — all free, no login required.
Try the toolsThe "How AI Reads You" course and the Blackbird AI Audit — from understanding patterns to applying them professionally.
Explore trainingSector Expertise
Financial services, legal, healthcare, education, enterprise, government, insurance, consulting, pharmaceutical, and more. Each with compliance maps, tool recommendations, and workflows.
Explore the AI Field Guide →Independence
Not funded by any AI company. Not affiliated with any provider. Not sponsored by any vendor. The research exists because the patterns exist.
Start Here
Whether you're a researcher, a professional using AI for consequential decisions, or someone who noticed the AI agreeing with you too readily.