The UK's NHS says it will anonymize data given to Google's DeepMind that is used to analyze blood test results and detect risk of serious medical problems
Timothy J. Seppala / Engadget :
Context & Ripple Effects
This announcement lands two years into a running dispute over DeepMind's access to NHS records. A 2016 report found Google had been given broader access to patient data than publicly announced, and in 2017 the National Data Guardian concluded that the transfer of 1.6M Royal Free patients' records rested on an inappropriate legal basis.
Committing to anonymize the blood-test data stream is the NHS's structural answer to that criticism: rather than defending the original transfer, it changes the form of the data going forward. The move also echoes the earlier eye-scan project, where DeepMind worked from anonymized scans.
First-order effects
- DeepMind's blood-test analysis work continues, but on an anonymized feed — removing identifiable patient records from what Google receives.
- The NHS trust behind the arrangement reduces its exposure to a repeat of the National Data Guardian's finding that prior sharing lacked a proper legal basis.
Second-order effects
- Other NHS bodies negotiating with AI vendors now face pressure to write anonymization commitments into their own data-sharing agreements, since the unanonymized version of this deal became the case study in what not to do.
- Google gains a compliance template it can point to in future health-data partnerships, shifting the argument from whether tech firms should get NHS data to under what data format.
Third-order effects
- If the pattern holds, UK public-health data deals converge on anonymize-first structures as standard, with identifiable transfers becoming the exception that invites scrutiny or litigation — as later seen when a court ultimately threw out the case over the original 1.6M-record transfer.
- The longer-term effect is that the burden of proving data minimization shifts onto the tech partner up front, making anonymization capability a competitive requirement for AI firms seeking health-system contracts.
The trend: Public health systems are restructuring AI partnerships around anonymized data feeds and explicit legal bases, converting ad-hoc record transfers into audited, minimized disclosures.