Craftdash: Building trustworthy cultural AI at startup scale

This case study explores how Craftdash Limited is developing Culture Illustro, an AI-powered platform designed to generate culturally accurate, high-trust visual content for use in education, storytelling and design. Built within the constraints of a small UK-based startup, the project focuses on embedding governance, provenance and cultural integrity directly into the data and development lifecycle.

Participation in Bridge AI – The Turing Way Practitioners Hub has supported this work by strengthening Craftdash’s use of systems thinking, stakeholder mapping, data readiness frameworks and lightweight governance-by-design. Together, these approaches enable the team to build responsible AI systems without the heavy operational burden typically associated with enterprise-scale governance.

This case study is published under The Turing Way Practitioners Hub 2025-26 Cohort - case study series. The Practitioners Hub is The Turing Way project that works with experts from partnering organisations to promote data science best practices.

Key takeaways

• Non-Western and underrepresented cultures are particularly at risk of inaccurate representation through generative AI models.

• Cultural integrity must be built into AI systems at the data and workflow level, not added later.

• High-quality, context-rich datasets are essential for reducing misrepresentation.

• Governance-by-design can be implemented at startup scale through lightweight, repeatable processes.

• Systems thinking helps identify where risks originate and where safeguards are most effective.

• Small teams can build responsible AI by prioritising key control points.

• Collaboration with cultural contributors and stakeholders is critical to maintaining trust and accuracy.