In a world of information abundance, Project ECHO aims to use AI to transform knowledge into wisdom through contextualized support. The ECHO Model has always been about people learning from people; AI is a tool that helps that learning reach further, not a substitute for it.
We at Project ECHO acknowledge our responsibility to develop systems ethically that create value for society by addressing concerns about Accountability, Transparency, Explainability, Privacy, and Security. In the process, we ensure the agency of the users while they interact with the platform.
Our commitments
- AI augments human connection, never replaces it. Human expertise and validation remain central to the ECHO Model. Subject matter experts, hub teams and participants stay in the loop wherever AI is used.
- Responsible, transparent and in service of communities. ECHO is committed to using AI responsibly, transparently and in service of the communities we support. We are open about where AI is used on iECHO and what it does.
- Greater participation and inclusion. AI can help overcome language and access barriers, enabling greater participation and inclusion in resource-limited settings.
Five key principles for responsible AI governance
These principles guide every decision about how AI is designed, trained and deployed across the ECHO community:
| 1. Open, interoperable public goods | Ensuring strategic autonomy for the community |
| 2. Diverse and representative data | Grounded in localized expertise from communities, reducing bias |
| 3. Multimodal and multilingual | Eliminating barriers to participation, ensuring each voice is heard and valued |
| 4. Human validation and expert curation | Data grounded in real insights, with humans in the loop |
| 5. Trust and safety architecture | Privacy and security at scale through de-identification |
How we are implementing the 5 governing principles
- Open, interoperable public goods
AI capabilities are built as digital public goods that work with open standards, so the community keeps strategic autonomy over the tools it depends on rather than being locked into any single vendor. - Diverse and representative data
Models are informed by localized expertise from the communities ECHO serves, across sectors, geographies and languages, which reduces bias and keeps guidance relevant to local context. - Multimodal and multilingual
Text, voice and video, in the languages people actually speak, so that language, literacy and connectivity are not barriers to taking part. - Human validation and expert curation
AI outputs are grounded in real insights and reviewed by people, while ensuring traceability to trusted knowledge sources. Experts curate what is shared, and humans remain accountable for decisions. - Trust and safety architecture
Privacy and security are designed in from the start, with de-identification so that AI can operate at scale without exposing personal information. See Privacy and Security for the controls behind this.
We are in process of developing the full fledged integration of AI features on iECHO. This rollout will be done gradually over the next year, in consultation with partners. Partners can choose to opt out of AI features at any time. No data will be used for AI model training
Questions about AI on iECHO?
If you would like to know more about where AI is used on the platform or how it is governed, write to ai@iecho.org.