BiNet USA / Current work
Research and Evidence
Sources before claims. People before extraction.
Research should make a question clearer, not turn a draft into a result or a person's testimony into someone else's property.
Separate the kinds of knowledge
A public record, private evidence item, personal account, proposed reform, and AI-generated summary are not the same thing. Each needs its own date, source, status, and permission boundary.
Statistics need a named dataset, publication date, population, definition, and limitations. This release does not publish new health statistics, survey findings, or clinical conclusions.
Consent and authorship
Participation in research is not automatic consent to publicity, advertising, future training data, or political use. Agree on the permitted use, confidentiality, attribution, withdrawal limits, and compensation before collecting sensitive material.
Private source files are not public simply because their ideas helped inform a policy page. Raw workbooks, credentials, court records, medical information, family records, and private correspondence remain excluded.
AI and evidence preservation
Use AI to assist organization and drafting with human review and source references. Do not generate a quotation, imply a finding, or invent a date to fill an archival gap.
Keep original evidence distinct from redacted derivatives and summaries. A revision should leave a trace of the change, not silently substitute a new account.
Commission research
A proposal should identify the question, jurisdiction, sources, intended use, affected-community review, deliverables, and budget. No study enrolment or sensitive testimony-upload service is opened here.