Applications for Artificial Intelligence are becoming more widespread than ever. Long past are the days when AI simply meant a really smart assistant on your phone who could do a few chores for you. Now, we have an AI-enhanced microscope that can diagnose deadly blood infections.
The invention comes from microbiologists from the Beth Israel Deaconess Medical Center (BIDMC), who hoped to create something that would compensate for the shortage of highly trained microbiologists.
The team at the BIDMC trained a convolutional neural network, which models the visual cortex in mammals in order to analyze visual data. Some 100,000 images were fed to the AI from 25,000 slides that had been treated with dye ahead of the analysis in order to make the bacteria more visible. Before long, the AI learned to sort the bacteria based on visible traits, achieving an accuracy of nearly 95%. When no dye was used, the accuracy levels dropped slightly to 93%.
While the numbers are already excellent, more training could result in a machine that is extremely reliable and could be used to save time for those who don’t have much of it, saving lives in the process.
Use of Artificial Intelligence in the medical field is only in its incipient phases, but there have been numerous experts who expressed a belief that in the future AI could, at the very least, help with setting diagnostics by looking at all possible options given the doctors’ observations, blood work, health history, and so on. An AI could very well draw up a few options from which doctors could pick from, and even help set a proper treatment plan.
Not everyone is happy when they hear about Artificial Intelligence, thinking that it spells out our doom as race on the Planet the day AI reaches Singularity. The truth of the matter is that, for now, AI is far from being a threat and it can be quite useful in day-to-day issues, including medical ones.
AI can also be helpful in helping doctors quickly access a comprehensive collection of data on patients, including from smart machines such as the one invented by the BIDMC, which are surely going to be created in the future. At this point, it’s just a matter of time before more and more ingenious uses are found for AI.
Like any new technology, Artificial Intelligence a society acceptance life-cycle. In the beginning, only a small group will embrace it, using AI for a matured solution. When the benefits start showing, whether we’re talking about someone’s life being saved, or some kind of improvement brought to our daily lives, others start accepting it as well. As history proves, the process of accepting AI’s going to have a peak of inflated expectations, then a trough of disillusionment, finally a slope of enlightenment as hype curve works. Eventually, we as end users, are humans who have a nature that influences the world.