Consequential AI governance
Keep life first.
A practical framework for teams that need to protect life and dignity, preserve agency, make proportionate decisions, assign accountable humans, and repair harm when systems fail.
Public working draft v0.3
Life First is not certification, legal advice, an emergency protocol, or a substitute for domain expertise. Stronger applicable rules still apply.
Start here
Follow the framework from principle to practice.
Choose a path below to review the framework, requirements, implementation, records, examples, or governance.
Purpose and commitments
Purpose, commitments, risk tiers, safeguards, and the operating cycle.
Explore the framework →Requirements and safeguards
Human authority, evidence, agency, pause conditions, records, and repair.
See the requirements →Apply the framework
Review gates, implementation phases, an eight-week pilot, and measurement.
Open implementation →Records and evidence
Intake, impact, evidence, decision, control, metric, appeal, and repair records.
Open the toolkit →Practice across settings
Fictional cases across community, health, belief, security, machinery, and extreme environments.
Explore the casebook →Stewardship and governance
Independent stewardship, provenance, governance, contribution, and public boundaries.
Read Governance & About →Connection to The Wizard Nexus
Independent stewardship. Clear portfolio relationship.
Life First is independently stewarded by Roshi. The Wizard Nexus references it as a human-first governance framework for consequential workflows; that connection does not transfer ownership or governance of Life First to TWiN.