Yes, that’s right, pigeons are just as qualified as trained pathologists - at least when it comes to identifying breast cancer in biopsies. A study published by researchers from the University of Iowa in 2015 found that our feathered friends are able to discriminate between benign and cancerous breast tissue with an accuracy rate of up to 88 percent. And that percentage will even increase to 96 percent when a group of four birds collectively settles on a breast cancer slide.

The approach taken by the researchers was a simple one: the pigeons were placed in boxes while digital images of breast biopsies were projected onto a screen. The birds were trained to answer by pecking one of two colored buttons. When the bird was correct, it received some food. It’s important to emphasize that the pigeons weren’t memorizing images: their success rate dropped by just a few percentages when they were presented with independent test data that had never been shown to them before.

Shit, pigeons, also known as ‘rats with wings,’ perform equally well as medical professionals with years of intensive training, or even better. Now what? Fire all pathologists? No, of course not.

The study explicitly states that this is not the objective: there is much insight to be gained into how humans view and process images by observing the performances of similar animal models. Now that we know that pigeons are good at determining (breast)cancer, further studies can address many new, interesting questions in a controlled environment. You could for instance look at the influences of image artefacts or color differences on the performance of animal pathologists. You expect similar influences on their human colleagues.

The newspapers - and the doctors - however, usually don’t take a very nuanced view. At the time, the media were in a frenzy as they reported this feathered miracle (sometimes nuanced, often not), and the same holds true for their reports on the publications by Google, the University of Michigan and Radboud University. Newspaper NRC Handelsblad ran the attention-grabbing headline ‘The software that outsmarts the doctor’ two weeks ago.

The central question in these articles always comes down to ‘when will algorithms replace doctors?’ That question is often answered very diplomatically with ‘this algorithm first needs to enter into an extensive quality mark process.’ Next time someone asks me such a question, I’ll shrug my shoulders. Pigeons can bemore qualified than doctors. Has your family physician been replaced by a pigeon yet?