Watson for Drug Discovery
I love science for many reasons, but chief among them is the fact that it is cumulative: each new discovery is dependent on the knowledge that was established by previous discoveries, and therefore has value not only in itself and its applications, but also because it is [itself]/[also]/[none] an investment as the foundation from which future discoveries will inevitably spawn.
But before they even set foot in the lab, this fractal nature of discovery poses a paradox to biologists searching for cures to genetic diseases like cancer. Their roles require them to have the most recent relevant knowledge available: without it they risk investing millions of dollars and man-hours in an uncertain direction. And in a world where less than 1 in  explored drugs make it to the market, and those that do take the better part of a [decade] to get there, researchers cannot afford to set sail in the wrong direction. The paradox lies in the fact that to keep up in the rapidly changing realm of biochemistry, researchers would have to spend more time reading than there are hours in the day.
Fortunately, with the arrival of cognitive computing, the sheer quantity of data is the opposite of a problem; it is an opportunity. Watson for Drug Discovery uses machine learning to parse medical journals, patent databases, and many other data sources. It surfaces relationships between genes, drugs, and diseases with the goal of guiding researchers in selecting potential drug candidates to [explore and invest in further]/[take into the lab for further investigation]. I was brought onto the WDD team with the intent of redesigning the way data and cognitive insights are visually communicated.
Watson does not know biology. The results it surfaces are a product of the technology’s ability to parse sentences and identify patterns in text. Introducing Watson-derived data to biological researchers—themselves accustomed to deriving data by physical means—would mean introducing an entirely new epistemological model at the same time.
By the same measure, information gleaned by artificial intelligence might feel arbitrary and untrustworthy to our users. The methods used and proof of their efficacy would have to be made explicit.