Wait, what? How do search results determine lung cancer? Basically since lung cancer doesn’t suddenly appear in one’s body, there are signs leading up to it that could be clues. To try and do this, they went backwards based on the user’s search. For example someone searching for “I have been diagnosed with lung cancer,” could have previously searched for symptoms that are related to it, such as chest pain, blood in sputum, bronchitis, and so on.
Microsoft’s algorithm will also take into account the person’s gender, race, age, and where they live, such as if they live in areas with high amounts of asbestos and radon, there could be a higher chance that they might have cancer. They also try to search for other indications, like whether they smoke by seeing if they have searched for products like Nicorette gum.
Is this accurate? According to Microsoft, this depends on how many false positives you are willing to put up with. If people see the message and choose to act on it, they have a chance of catching it early, if they don’t, then that’s the risk they’ll have to take. Of course Microsoft isn’t saying you should rely on them for a diagnosis, but if it could help potentially detect the symptoms early, why not?