Responsible AI & Medical Use

Responsible technology starts with clear boundaries

AI can meaningfully support scientific and healthcare work — but only with human oversight, documented intended use, and honesty about its limits.

1. Human oversight

High-impact outputs should be reviewed by appropriately qualified people before they inform a decision.

2. Intended use

Every system should have documented users, purpose, limits, and prohibited uses.

3. Data governance

Minimise data, establish a lawful basis, control access, define retention, and assess suppliers.

4. Evaluation

Test accuracy, robustness, bias, security, and operational fitness in the context of use.

5. Transparency

Tell users when AI is involved and explain material limitations plainly.

6. Regulatory status

No product is called a medical device or SaMD on this site unless the exact product and status are confirmed. "Designed for possible future regulated use" is not the same as approval.

Medical-use boundary

This website is not a medical service

Nothing on this website is medical advice, diagnosis, or treatment, and it must not be used for emergencies. If you need urgent medical help, contact 999 / 112 or your local emergency service.