Our Data Science & AI Commitments

At Sunstone Institute, data science and artificial intelligence are core research capabilities, not ends in themselves. The following commitments define how we design systems, govern data, and ensure that technological capability is always matched with scientific integrity and human judgment.

Open Science

We make our data, methods, code, and results as accessible, transparent, interoperable, and reusable as possible, so that others can verify findings, build upon them, and apply them for public benefit. Our approach to open science is informed by and aligned with the UNESCO Recommendation on Open Science (2021).

Selective Protection

We share everything we can and protect what we must. When ethical, legal, security, or community-governance concerns prevent disclosure, we explain exactly why. We follow FAIR (Findable, Accessible, Interoperable, Reusable) principles wherever possible.

Interoperability

We design systems so our data, methods, and insights can be reused, verified, and combined across tools, institutions, and borders. This preserves our integrity while maximising analytical and societal value.

Scalable Infrastructure

We build our platform to grow in capability and reach without compromising reliability, security, or scientific integrity. We use modular architecture, open standards, and strong governance. We ensure the platform remains robust, adaptable, and trustworthy as data volumes and use cases evolve.

Traceable Knowledge

We maintain a digital trail for every study, making our research reproducible and auditable. Anyone can review our sources and follow our steps to see exactly how we reached our conclusions – now and years from now.

Security Culture

We treat security as an organisational practice, not a compliance exercise. This means defence-in-depth across access management, encryption, secure development, operational security for sensitive research, and continuous risk assessment.

Data Stewardship

We apply robust data governance across the entire data lifecycle—from collection to reuse. This includes ethical data collection and use, clear ownership and licensing, defined stewardship responsibilities, and accountability mechanisms for all data and AI practices.

Inclusive Knowledge

We actively reduce barriers to access and participation. We avoid extractive data practices and respect local, community, and Indigenous knowledge where relevant to our research.

Human-Led AI

We use Artificial Intelligence to support pattern discovery and analytical exploration, never to replace human judgment. All AI outputs undergo human review, and final responsibility rests with our researchers. We document model assumptions, limitations, and uncertainty, actively test for bias, and monitor AI systems over time.

10.

Purposeful Automation

We automate repetitive and error-prone tasks so our team can focus on asking good questions and interpreting results. We apply software and data engineering best practices – automated quality checks, robust metadata management – but we do not pursue efficiency at the cost of human interpretation, contextual understanding, or ethical judgment.