Mission and objectivesMission To create a unique nexus where civil society meets and engages the academy, industry, and government to collectively mitigate risks and make smarter decisions toward more resilient, democratic, and prosperous communities. As a centre of excellence, GCRI excels in research, innovation, and capacity building across enterprise risk and innovation management, addressing the societal impacts of technological disruptions and financial evolution. Uniting experts from economics, finance, policy, and technology, we are committed to building interdisciplinary tools, capacities, and communities for risk mitigation, resilience building, and sustainable development Objectives GCRI's objectives include: 1. Global Knowledge Exchange: Creating a platform for interdisciplinary discussion and research on risk and resilience, fostering global collaboration among industry leaders, academics, policymakers, and practitioners 2. Standardization and Metrology: Developing and promoting standards and measurement practices to enhance the reliability and interoperability of risk management strategies. 3. Systems Innovation: Building tools, capacities and communities for the development and application of risk management technologies, standards, and innovation ecosystems. 4. Scholarly and Industry Publications: Publishing standards, reports, and scholarly articles to contribute to academic and practical discourse in risk management. 5. Liaison with Global Entities: Engaging with civil society and maintaining technical interactions with international entities to promote a unified approach to risk management challenges. 6. Advanced Study: Investigating fundamental and applied domains related to risk management, including natural and human-induced risks, and developing advanced tools for risk intelligence and emergency management
ContextArtificial Intelligence is becoming central to how societies forecast, respond to, and recover from disasters and climate-related crises. Yet, as the scale and speed of AI deployment increase, so does the risk of opaque models, unverifiable outputs, and decision-making processes that lack transparencyโespecially in high-stakes domains like disaster risk reduction (DRR), climate adaptation, humanitarian logistics, and infrastructure resilience. To ensure that AI serves humanityโs most urgent needsโethically, reliably, and responsiblyโwe must build a new kind of infrastructure: one that prioritizes verifiability, reproducibility, and trust in every aspect of the AI pipeline. Vision: We are forming a global open-source community under the Observatory Protocol (OP)โan initiative led by the Global Centre for Risk and Innovation (GCRI)โto develop the worldโs most robust AI verification system dedicated to planetary-scale risk intelligence and climate action. This effort will: Unite researchers, engineers, and institutions across continents to co-develop a new standard for verifiable AIโfrom data provenance to inference results. Establish a decentralized, permissionless environment where any AI model used in public decision-making can be reviewed, tested, and certified for integrity and reproducibility. Build a transparent trust layer for AI models in DRR, sustainability, and climate forecastingโensuring outputs are auditable, tamper-proof, and scientifically sound. Foster a public good ecosystem of verifiable computation, enabling equitable access to validated models for all governments, researchers, and humanitarian actors. Strategic Outcomes: A global registry of verified AI models for disaster and climate resilience A shared testbed for model reproducibility and verification Protocols for cryptographic logging, metadata auditing, and model traceability Open community guidelines for ethical, transparent, and explainable AI deployment Foundations for policy, insurance, and infrastructure decisions built on trusted machine intelligence
Task DescriptionWe are seeking 50 Online Volunteers to support the weekly review and verification of open-source AI models as part of our Trusted Inference initiative. Under the guidance of the technical focal point and in collaboration with the core development team, the Online Volunteer(s) will contribute to ensuring the integrity, reproducibility, and transparency of AI model outputs. Specifically, each volunteer will be responsible for: Conducting verification testing on at least one AI model using a standard checklist; Documenting and comparing test results across different systems and environments; Summarizing key metadata, including model inputs, architecture, dependencies, and runtime settings; Recommending improvements to enhance model reproducibility or output verifiability; Contributing test outcomes to a shared GitHub-based \"Verified Model Registry\" or review log. Work will be conducted remotely using collaborative tools such as GitHub, shared documents, and virtual check-ins. ๐ฆ Key Deliverables Verification checklist applied to at least one open-source AI model Documented test results showing consistency across systems and environments Metadata summary for model input, architecture, dependencies, and runtime context Suggested improvements for enhancing reproducibility or verifiability Contribution to a shared \"Verified Model Registry\" or review log on GitHub โ Success Outcome One or more AI models reviewed, tested, and added to the OP trust layer Clear evidence of reproducibility and output integrity validated by an independent team member Documentation published publicly for use by other contributors in the verifiable AI community Contribution acknowledged as part of the OP Verified AI Library and global working group roadmap
Competencies and values
Living conditions and remarks