Alphabet president Ruth Porat says Google’s AI work is being used in cancer detection and diabetic-eye screening, putting concrete healthcare claims behind the company’s enormous AI ambitions. The figures are notable, though they need to be read as claims reported by Fortune, not as a new independent evaluation.
Watch Desk analysis
What happened
Speaking at Fortune’s Most Powerful Women Summit, Porat said a 2020 study found Google’s mammogram-reading AI reduced false negatives by 9.4 percentage points in the US and 2.7 points in the UK, compared with radiologists. She also said Google and local partners had completed about one million diabetic-retinopathy screening scans, with a target of more than six million.
Porat connected those examples to Google’s planned $200 billion AI buildout and argued that early detection could help widen access to healthcare. The Fortune report attributes the figures to Porat and references the 2020 study; the excerpt does not give the study’s methods or details of the screening programme.
Why it matters
These are different kinds of evidence. The mammogram figures refer to a study comparison, while the scan total describes the scale of a screening effort. Neither, by itself, tells us how well the systems perform across populations, how many people receive follow-up care, or whether screening improves health outcomes.
Still, the examples show what a vast AI investment is meant to support beyond data centres and model launches: tools intended to help detect disease earlier. That promise matters most if it reaches patients and clinicians who have been underserved, not simply if the scan counter keeps turning.
Our read
The figures make for a more useful AI story than another abstract spending forecast. But “more scans” and “fewer false negatives” are starting points, not the whole measure of success. The next useful evidence is how these tools perform in practice, for whom, and what happens after they flag a possible problem.
What to watch
- Whether Google or the study’s authors provide the methods and population details behind the mammogram comparison.
- Whether the diabetic-retinopathy programme reaches its stated goal of more than six million scans.
- What outcomes and access measures are reported alongside screening totals.
Discussion spark: For AI screening tools, what should count most as proof of success: better detection, wider access, or improved outcomes for patients?
Sources and evidence
- Source update (13 October 2026, 14:52 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.