Post #1772990
2026-04-23 19:31 UTC
Contrast this with Ben Goldacre's OpenSafely which was specifically designed so that researchers never got access to individual's data. Instead they submitted "questions" (with code) that was peer reviewed (in the open) with explicit checks to make sure no PII came out in the answers. Then the code ran inside the data centre and the answers returned. By design it put technical and structural breaks between researchers and PII to avoid these sorts of problem being able to occur.
Replies (2)
-
@slowe@mastodon.me.uk 2026-04-23 19:35
Here's the Register reporting on the UK Biobank story https://www.theregister.com/2026/04/23/500k_biobank_volunteers_data_listed/
-
@otfrom@functional.cafe 2026-04-23 20:50
@slowe there are so many areas where I'd love this style of sending code to the data