Xiaowu Gai

I've spent my career in clinical genomics — sequencing genomes, interpreting variants, solving cases, running expert panels. For years, something bothered me that I couldn't quite name. In 2021, I finally gave it a name: Genomic Dignity. Here's what I mean - and why it matters.

If you're building genomic data infrastructure and you're thinking about how to operationalize data governance: these approaches exist and they work. Genomic Dignity is something we can engineer into the system. We don't have to wait for policy to catch up.

I'm currently pursuing further work on AI-assisted approaches to genomic variant classification and privacy — building on the same principle: that the tools we build for genomic medicine should reflect the dignity of the people whose genomes we're working with.

What We Built: Encryption and Watermarking for Genomic Data [The technical solution — from concept to patents]

Genomic research will be more powerful if participants trust the institutions that hold their data. That trust is built best through dynamic consent, not through language designed to maximize participation. Genomic Dignity isn't a precondition for science that lasts.

The practical barriers are real: dynamic consent requires infrastructure, communication systems, and ongoing engagement that most research programs don't currently have. But the barrier isn't technical feasibility. It's institutional will and research funding priorities.

Whose Data Is It, Really? Where Genomic Consent Fails [The consent crisis — what patients aren't told]