About Etika
Trust in AI cannot be self-declared.
Etika is an independent certification authority that does for AI what ISO 20000 did for IT services and what the FSC did for forestry: it provides an independent, certification-backed mark that an organisation meets a published, public standard.
Mission
To make ethical, climate-aware, and labour-respecting AI development verifiable — not merely claimed in marketing or pledged in voluntary charters.
The AI industry has produced an unprecedented amount of self-issued ethics frameworks. Almost none of them are independently certified. Etika exists to close that gap.
Method
- 01
Application
An organisation submits a scoped application identifying the AI systems to be assessed. Etika assessors review every submission personally.
- 02
Documentation review
Etika assessors examine model cards, data sheets, impact assessments, governance records, and emissions reports.
- 03
On-site & technical certification
Independent assessors conduct interviews, reproduce evaluations, inspect labour conditions, and verify monitoring controls.
- 04
Ethics review
Every certified system passes through Etika's independent Ethics Review Council.
- 05
Decision & publication
If granted, the Blue Ribbon and certificate appear in the public registry. Non-conformities are classified and tracked publicly.
- 06
Continuous integrity programme
Unannounced surveillance visits, a public complaints register, and a recurring Ethics & Safety Culture review.
Independence
Etika is incorporated under Canadian law and headquartered in Toronto, Ontario. No board seat is held by an employee of any AI lab or platform company. Certification fees are structured to prevent any single organisation from accounting for a material share of our revenue.
All standards documents, certification methodologies, and decision rationales are published openly. The certification mark belongs to Etika, not the certified entity, and may be withdrawn for cause at any time.
Published certification methodology
Etika's certification procedures are published in full so that certified organisations, regulators, and the public can verify how a determination was reached.
- AM-01Read →
Scoping & Certification Planning
Procedure for delineating in-scope AI systems, mapping stakeholders, and setting the certification plan.
- AM-02Read →
Documentary Evidence Review
Required artifacts and evaluation rubric for completeness and credibility.
- AM-03Read →
Technical Reproduction Protocol
How assessors independently reproduce key model evaluations.
- AM-04Read →
Worker & Contractor Interviews
Structured interview protocol with psychological-safety provisions.
- AM-05Read →
Ethics Council Referral
Mandatory referral of every certified system to the independent Ethics Review Council.
- AM-06Read →
Non-Conformity Classification
Definitions and remediation timelines for minor, major, and critical findings.
- AM-07Read →
Unannounced Surveillance Cadence
Post-certification evidence schedule and re-certification thresholds.
Methodology suite v1.0 · last reviewed 2026-06