A one Ecosystem Initiative

Implementing OECD AI Principles for Everyone

An open source community translating globally-adopted AI ethics standards into practical, auditable tools. Built on the one Framework.

OECD AI Principles

01

Inclusive Growth, Sustainable Development and Well-being

AI should benefit people and the planet, not just shareholders.

02

Human Rights and Democratic Values

AI must respect privacy, fairness, and human dignity.

03

Transparency and Explainability

People should understand how AI systems work and affect them.

04

Robustness, Security and Safety

AI systems must be secure, safe, and resilient.

05

Accountability

Those who develop and deploy AI must be accountable for its outcomes.

46
Countries
G20
Endorsed
2024
Updated
5
Principles

The Eight Golden Rules

1 Your Data, Your Device Privacy

AI processes your data locally by default. Cloud requires explicit consent.

2 No Training Without Consent Privacy

Your work is never used to train AI models without explicit permission.

3 Transparent Data Flow Transparency

Users always know where their data goes and who processes it.

4 Freedom From Lock-In Inclusive

Export your data, switch providers, and self-host without barriers.

5 Explicable Decisions Explainability

Understand why AI decides what it does and who is accountable.

6 Security By Default Safety

Strong encryption and minimal attack surface protect user data.

7 Honest About Limitations Accountability

AI systems communicate what they can and cannot do.

8 Compute For Good Well-being

Idle compute should contribute to humanitarian causes.

Open Source

Compute For Good

Your GPU should save lives, not sit idle.

Join the EAF compute team and contribute to humanitarian research.

Join Team

Public Roadmap

Open Source Foundation Q1-Q2 2026
  • Launch certification criteria v1.0
  • Release self-assessment tool
  • Establish compute donation team
  • Onboard first 10 community auditors
Community Growth Q3-Q4 2026
  • Partner with EFF, Privacy International
  • Certify 10 early adopters
  • Launch badge program
  • Publish compute contribution metrics
Ecosystem Scale 2027
  • Decentralized audit network
  • Multi-language support
  • EU AI Act integration
  • 1 petaFLOP/day compute goal

The one Framework

O

Ownership

Your data belongs to you. Local-first by default, cloud only with consent.

N

No Lock-In

Export your data, switch providers, and self-host without barriers.

E

Explicability

Know how AI decides and who is responsible for its outcomes.

Privacy-first. No vendor lock-in. Security by default. Simplicity over complexity.

Learn about one

Join the Community

Help build the open source implementation layer for international AI ethics.