About Ink Stone
Scholarly in Outlook.
Deliberate in Practice.
We are a small Hong Kong practice working at the point where AI capability meets institutional knowledge work. Our clients are organisations with long-standing obligations to accuracy, integrity, and public trust.
← Back to HomeOur Story
Started from a specific frustration
Ink Stone was founded in Hong Kong in 2022 by a small group of practitioners who had spent their careers in policy research, academic publishing, and heritage management. What brought them together was a shared unease about the way AI tools were being offered to institutions like theirs — as generalised products, with implementation guides designed for commercial contexts, and no meaningful understanding of what it means to maintain standards in a scholarly or curatorial environment.
The name comes from the inkstone — the grinding surface on which ink is prepared before it meets paper. It is a tool of quiet, preparatory work, valued for the consistency and quality it enables rather than for any effect of its own. That is how we think about AI in institutional settings: as a prepared surface, not a finished product.
Our practice is deliberately narrow. We work with three types of institution — research and policy centres, book and journal publishers, and cultural organisations such as museums, libraries, and heritage foundations. We do not try to serve every sector, and we are not trying to scale quickly. The organisations that work with us tend to value that restraint.
Our Mission
Enable, without displacing
Our mission is to help knowledge institutions in Hong Kong and the wider region use AI in ways that are genuinely useful, proportionate to their standards, and honest about what AI can and cannot do. We take seriously the idea that institutional trust is hard-won and easily damaged, and we design our services accordingly.
Every engagement we take on involves qualified human review at meaningful points in the workflow. We do not sell automation as a replacement for professional judgement. We sell careful AI support as a way of freeing qualified people to apply their judgement where it matters most.
Our Values
- —Restraint: We recommend only what an institution actually needs.
- —Transparency: We are clear about what AI produces and what it cannot judge.
- —Calibration: We build our approach around your existing criteria, not generic templates.
- —Continuity: We stay involved across the review cycle, not just at setup.
The Team
People Who Have Done This Work
Our team includes practitioners from policy research, editorial work, archival management, and applied AI. We do not separate the people who understand the institutions from the people who understand the technology.
Wai-Lin Cheung
Founding Director
Fifteen years in policy research across government-adjacent institutes in Hong Kong and Singapore, with subsequent advisory work on knowledge management and information governance.
Marcus Ho
Editorial Practice Lead
Former commissioning editor at an academic press in Hong Kong, with ten years of experience evaluating manuscripts across humanities, social science, and public policy titles.
Siu-Yung Lam
Cultural Institutions Lead
Spent twelve years in curatorial roles at two major Hong Kong cultural institutions before moving into consultancy on digital collections strategy and public engagement programming.
Standards
How We Maintain Quality
Human Review at Every Output
No AI-produced output reaches your staff or public without passing through a qualified human reviewer. This is a design principle, not a safety disclaimer.
Written Criteria Agreements
Before we deploy anything, we document your criteria in writing and review them with you. These criteria documents are living records, updated at each quarterly review.
Data Confidentiality
Institutional materials are treated as confidential. We do not use client data for model training or sharing. Confidentiality terms are included in every engagement agreement.
Quarterly Calibration Reviews
Every service includes quarterly reviews at which we assess whether the AI outputs continue to meet criteria, what has changed in your work, and what adjustments are warranted.
Ethics Review for Public-Facing Work
Cultural institutions receive a structured ethics review at the outset of their engagement, covering representation, provenance, appropriate use, and the standards expected of their public audiences.
Engagement Transparency
We do not obscure what AI is doing in your workflows. Where AI-produced content reaches staff, it is identified as such. We believe clarity about process is part of institutional integrity.
AI Integration in Hong Kong's Knowledge Sector
Hong Kong has a distinctive concentration of research institutes, policy centres, academic presses, and cultural foundations whose work depends on the careful handling of information — its accuracy, its provenance, its readability for different audiences. These institutions carry responsibilities that commercial organisations do not, and they operate under scrutiny that rewards consistency rather than speed.
Ink Stone works within this landscape because we know it well. Our practice was built by people who spent careers inside these institutions, not by technologists who later developed an interest in them. That difference shapes everything about how we structure our services — the emphasis on calibration, the insistence on human review, the quarterly rhythm that keeps AI support aligned with evolving institutional needs.
We are based at 8 Hollywood Road in Central, and we are open to initial conversations with institutions across Hong Kong and the broader region. There is no obligation attached to a first conversation, and we are candid about the cases where our services are unlikely to be a strong fit.
Interested in how we work?
An initial conversation is the right starting point. We will ask questions, listen carefully, and be straightforward about whether we can be of genuine use to your institution.
Start a Conversation