Today we published the Antrex AI Ethics Charter — the commitments that already govern our internal work, now open for anyone to adapt. It is short on purpose.
The five commitments
1. Cite or abstain — no confident answers without traceable sources. 2. Privacy is architectural — on-device by default, no advertising profiles, ever. 3. Human precedence — in domains of faith, health and law, AI points to qualified humans at the boundary. 4. Explain the machine — users can always ask how an answer was produced. 5. Measure harm, not just accuracy — every release ships with a failure audit.
“Ethics documents that change nothing are decoration. Ours ships in our release checklist.”