Role of AI in Modern Medicine
By Ali Reza
The way researchers diagnose and treat diseases has changed over time—and it seems poised to change again. In fact, new experiments point to a possible use of artificial intelligence in the medical research field. If deployed properly, these AI units could effectively revolutionize the search for better, stronger, more effective medications.
The Way it Was
The search for drugs has not always been a particularly efficient one. Historically, as researchers have tried to diagnose and treat diseases, they have devoted their search to isolating the single mutation on a given gene that might have caused the problem. The problem with this approach is that the onset of a disease is typically triggered by more than just a single genetic variant. There is usually a complex array of factors in play—environmental issues, multi-gene interactions, demographics, and more.
That’s obviously a lot of information to synthesize, which is precisely why drug research has been so laborious, at times so slow and so ineffectual. Computers have never been up to the challenge of analyzing all of this information, either, because there is simply too much of it to fit within standard databases.
Rise of the Robots
Where traditional computing has failed, artificial intelligence may hold some answers. Artificial intelligence may make it possible to tease out interactions from more complex and far-reaching health data, perhaps even offering a quick way to sequence full genomes and ultimately gather more molecular information than ever.
But what exactly does all this mean? What it means, most promisingly, is that precision medicine may one day be a real possibility, not just a pipe dream—and we’ll have AI to thank for it. AI may lead to a world in which researchers can identify all of the unique characteristics that a person has that might lead to the onset of a certain disease, and also to potential avenues of precision treatment.
The concept of precision medicine is nothing new. It has long been a sort of Holy Grail among medicine researchers. Actually treating human beings in a precise manner can be complicated though—because, well, human beings themselves are so complicated.
Precision medicine embraces that complexity. The concept behind it: All of us are wildly distinct, genetically speaking. We differ not only in our genetics, but also in our environmental histories. In precision medicine, the goal is to provide each person with a specific, individualized treatment that takes these distinctions into account—as opposed to the more generalized approach to treatment that is necessary today.
What the Research Says
But back to the robots. Researchers recently presented their approach to AI-empowered healthcare at a New York Academy of Sciences conference. What they presented was, basically, a scenario in which computers are asked to think about genomics, diseases, and disease treatments in the same way that human beings do—only much faster and on a grander scale.
The applications for such a scenario could prove myriad. One particularly noteworthy application is drug research. According to this initial report, AI could be used to identify new uses for existing drugs—uses that previous research methods may have missed. It currently takes an average of 14 years and $2.6 billion to develop a single drug for clinical use, so the implications of AI-enabled methodologies—for pharmaceutical companies as well as for patients—are obviously exciting.
In fact, the researchers who presented these findings have already used an AI system—they call it EMERGENT—to identify new biomarkers that could prove to be drug targets for glaucoma.
So what is the next step for all of this? One immediate goal is to develop a new method to visualize the data produced by these research robots. After all, the AI results will not be useful unless biologists can actually sit down, understand them, and extrapolate. Right now, apps are being developed to make this possible, including some that will allow researchers to immerse themselves in AI data within the confines of a gaming system!
All of it seems promising—and certainly, medical research seems like it could be the best application yet for artificial intelligence.