Lab 4 · Data governance · about 15 minutes
Can This Data Cross the Border?
AI in biology runs on data, and more and more rules decide where that data may go. A Seattle university and a partner university in Shenzhen want to run a joint AI-for-biology course. For each request below, decide: share it, get a review first, or don't share it.
- Recognize which kinds of data carry cross-border risk: human genetic data, health data, personal data, and data tied to restricted companies.
- See why "methods, not data" keeps international teaching collaborations simple.
- Know when to stop and ask your compliance or research office.
Key ideas · a 2-minute primer before you start
- Why data rules exist
- To protect individuals' privacy, to protect a country's interests in sensitive data such as its people's genomes, and to protect companies' intellectual property.
- Sensitive data types
- Human genetic and genomic data, health data, personal identifiers such as names, emails, and grades, and companies' confidential data. Public and synthetic data are generally safe to share.
- Where the rules come from
- National laws on each side, such as China's human genetic resources rules, its Personal Information Protection Law, and its Data Security Law, and in the US, HIPAA, FERPA, and the bulk sensitive data rule. Consent forms and contracts add their own limits.
- Minimum necessary
- Share only the data a task actually needs. Often that's a summary instead of raw records, or a course email instead of a full student profile.
- Share methods, not data
- Instead of moving sensitive data, send the code to where the data lives and share only the results. Federated learning applies the same idea to training AI models across sites without pooling the raw data.
- When to ask
- Bring in your compliance, research-ethics, or export-control office early whenever data comes from people, crosses a border, or involves a company.
Go deeper in the free course lesson Protecting Confidential and Patient Data.
This is not legal advice. The rules summarized here are simplified for teaching and change often. Real decisions belong with your institution's compliance, export-control, or research-integrity office.
Go deeper: the lesson Protecting Confidential and Patient Data in the free course AI for Life-Science Professionals covers which data can go into which AI tools, and why.
CityUx prototype · Content CC BY 4.0, code MIT. Scenarios are fictional and simplified; not legal advice.
