In the first case of AI (artificial intelligence) saving a patient's life, reported in Japan, AI correctly diagnosed a female patient with a rare type of leukemia, and did so in just 10 minutes by comparing the patient's genetic sequence and 20 million clinical oncology studies.
How will AI change the structure of our healthcare in the future?
Is big data in healthcare futuristic? Or is it present tense? What is the relationship between big data and artificial intelligence? There are a number of organizations internationally that study the application and development of big data in healthcare, leading engineers familiar with data algorithms to devise advanced algorithms that allow the results to learn and improve accuracy in a continuous loop. Often, computer algorithms take only a few minutes to help us sift through hundreds of millions of records to find the answer to an action, but it takes years for humans to accomplish the same element of the work. This is the benefit of combining big data with AI, and it suggests that AI will be the most important assistant and enabler of biomedical development in the future.
Biomedical leapfrogging began with the Human Genome Project
The Human Genome Project (HGP) is a project that aims to identify all the genes in the human race by collecting the DNA of people from all over the world, and mapping them out. The Human Genome Project has also spawned a number of development projects and human resources. In recent years, more than a million researchers have been engaged in studying DNA sequences across the globe, and research to explore the genetic basis of disease has been put on the agenda.
The human genome project has been in development for more than 13 years, and as computer hardware performance continues to improve, the amount of big data generated by the genome project has grown at an astonishingly high geometric rate, leading to the emergence of data-based precision medicine. In the past, it took the Welle-Sanger Institute in the UK 10 years to sequence DNA, but now it only takes an hour to do so.
The evolution of personalized medicine
One of the benefits of big data research in healthcare is the ability to accurately predict an individual's health and prevent disease. For example, by using a smartphone with a monitoring device, users can monitor their heart rate, distance traveled, calories burned, and so on. It's like having a doctor with you, giving you useful advice and alerting you when necessary. For example, it reminds you that your blood sugar is now above the alert level and that it's time to take an insulin injection.
All of this information can be analyzed and integrated into your personalized medical record, which can be securely stored in a "cloud database. Although in its early stages, these technologies are already being used in some places.
In the future, if you feel the need to go to the hospital, a doctor with an AI assistant might take out a tablet and look at your cloud-based medical record, which might include genetic sequences and other useful information, to confirm your diagnosis. Carefully incorporating AI assistants when counseling patients could increase the accuracy of doctors' diagnoses several times over.
Is there a risk to AI visits?
The risks of AI can be summarized into three main categories: program errors, cyberattacks, and misinterpretation of instructions. But all three of these risks can be avoided with proper planning.
Programming errors, commonly known as "bugs," are a common but avoidable reality in poorly developed software. Programming errors occur because the development and testing processes are not executed correctly. The consequences of program failures can be large or small. However, software programs have been used for decades in areas where security is critical, such as hospitals and the airline industry. It is reasonable to expect that what softwares can do before, medical AI apps can do as well.
Research in cybersecurity is well-funded, and it's still basically the devil's own business. We certainly shouldn't be overconfident that hacking is impossible, but there's no particular reason to think that medical AI can't protect against cyberattacks.
The shortcoming of listening to commands in the literal sense can be compensated for by internal protections, which are standard practice for other systems that require extreme security. There is no way that a hospital can have just one AI in charge of decisions that are critical to a patient's life or death, such as whether or not to shut down a vital system.
Despite the risks, we can manage medical AI with the same experience we have used in other fields over the past few decades, so why do we still need human doctors? Given the very real social and economic benefits that AI can bring, we should still be bold enough to explore the possibility of developing AI assistants, and of course, it must always be us humans who lead them!
Topics: Artificial Intelligence, Personalized Medicine, Big Data, What's New in Biomedicine, Precision Medicine, MedTech