Aletheia logo

About

Project Aletheia is an independent research lab building Basanos, an adversarial verification engine for AI-assisted science. It tests the engine against first-person reports, clinical material, instrument-linked records, and public scientific datasets.

The public research catalog currently contains 5 supported findings, 28 open questions, and 0 ready-now tests. No outcome currently has a verified-discovery label.

Why it matters: most research infrastructure is built to find supporting evidence. Aletheia is built to challenge proposed findings and keep failed predictions visible. The anomalous-science corpus is the stress test — a domain where selection bias, cultural priming, and wishful thinking usually win — and the deliverable is the discipline itself: a registry where nulls, controls, and evidence boundaries are first-class outputs.

Infrastructure

Aletheia is built locally and reproducibly. Public questions, tests, outcomes, and novelty reviews are kept separate from operational state. Every supported finding is gated by an operator decision before it can be cited.

Host systemLocal research workstation
Current top-level frameQuestion → test → outcome → novelty review
Public repogithub.com/0100001001101111/project-aletheia