Global Trust Challenge
A multi-year initiative focused on what it would mean for trust in AI systems to be justified rather than assumed.
What would an AI system have to demonstrate to be trusted?
Trust in AI systems is currently asserted, not demonstrated. Companies publish responsible AI principles. Institutions adopt AI policies. Auditors produce reports. But the underlying question what would a system actually have to show for trust to be justified remains unanswered in any rigorous, operational way.
The Global Trust Challenge addresses this directly. It brings together institutions across sectors to develop practical, evidence-based frameworks for justified trust: what evidence a system must produce, what it must document, and what pathways must exist for challenge and correction.
Frameworks for verified trustworthiness.
The Challenge operates across six sectors healthcare, education, public administration, journalism, financial services, and research developing sector-specific trust criteria grounded in the AI Commons infrastructure stack.
- Trust criteria frameworks what a system must demonstrate to meet justified trust thresholds in each sector
- Evidence standards what documentation, audit trails, and provenance records satisfy those criteria
- Procurement language model contract terms that institutions can use to require verified trust in AI procurement
- Pilot deployments working with partner institutions to implement and stress-test the frameworks in live contexts
- Public reporting annual findings on the state of AI trustworthiness across sectors
We are seeking institutional partners.
The Global Trust Challenge is most useful when it includes the institutions that will actually need to make trust decisions hospitals, universities, government agencies, newsrooms. If your institution is navigating AI procurement, deployment, or governance, we want to work with you.