AI’s Influence in Medicine
What is artificial intelligence?
From autonomous vehicles to portable devices capable of communicating with people on an international level to relevant web searches, technology in the modern era sparks with life. One thing the technological advancements mentioned have in common is that they used artificial intelligence. There are various forms of artificial intelligence. Humans all have intelligence, it allows them to think, process information, or problem solve equations. This is natural intelligence. Machines are distinct in that there is a different vocabulary associated with its ability to cognitively process information and data sets, which is relatively known as a machine’s artificial intelligence. But not all machines have artificial intelligence. For example, human-operated vehicles. Although vehicles are machines, it relies entirely on the human’s manual operation and cannot function on it’s own, unlike a self-driving car. While engineers work on a vehicle’s operation using AI, the medical field is also finding ways to implement artificial intelligence into the workplace. The medical field is extremely spacious in its categories, thus AI is being implemented easier in some branches of medicine than others. Some of the fields include histology, dermatology, and radiology. Artificial intelligence represents the future of technology and human efficiency. For medicine, artificial intelligence are arguably seedlings awaiting their transformation into a forest abundant with life. However, with such advancements exist setbacks and the fear of the unknown. How has AI been used in the field of medicine? What are some of the future possibilities concerning artificial intelligence in the field of medicine? What are some fears that people have towards machines having artificial intelligence?
AI’s Diverse Forms: ANNs, Fuzzy Expert Systems, and Evolutionary Computations
The field of medicine has implemented artificial intelligence into its technological applications since the 1950s. In the article “Artificial intelligence in medicine,” A N Ramesh from the Royal College of Surgeons of England introduces artificial intelligence and the various systems that have been developed or used in the field of medicine. The various forms mentioned include artificial neural networks, fuzzy expert systems, evolutionary computations, and hybrid intelligent systems which all have different functions and applications. For example, artificial neural networks, mentioned as “the most popular AI technique in medicine,” have been used in the “clinical diagnosis, image analysis in radiology and histopathology, data interpretation in the intensive care setting and waveform analysis.” In other words, artificial neural networks (ANNs) is being used to diagnose patients that come into clinics, verify and analyze images to help doctors pinpoint potentially cancer cells or lesions for radiologists and histopathologists, and also review a given data set in order to assist medical professionals that work in the intensive care department of medicine. With reference to the article, artificial intelligence appears to be extremely innovative. In further statistical detail, ProstAsure Index which is an algorithm that is derived from artificial neural networks created by Stamey BA, is capable of classifying prostates as benign or malignant. With a whopping diagnostic accuracy of 90%, ProstAsure Index would be extremely useful in helping a medical professional make a diagnostic for a patient. Generally, with the overview from a doctor, the diagnostic accuracy could detect what doctors couldn’t and vice versa, making the diagnostic process substantially more efficient and accurate. Moving onto the next form of artificial intelligence being used in the field of medicine as mentioned in the article by Ramesh is the fuzzy expert systems. “Fuzzy logic is the science of reasoning, thinking and inference that recognises and uses the real world phenomenon -- that everything is a matter of degree.” What this means is that the fuzzy logic peers into the natural world and understands that not everything is picture perfect. It looks at the “fuzzy” imperfections of reality. Fuzzy logic then handles medical applications that adopt the “if--then” rule. This rule, simply, has a simple “if” this variable exists in a given data set, “then” it means this patient could have the following outcomes. The fuzzy expert systems are more often used to diagnose cancers. Moving onto the next topic, evolutionary computation is a “general term for several computational techniques based on natural evolution process that imitates the mechanisms of natural selection.” In other words, an evolutionary computation system will take in separate datasets and develop itself on the basis of the solutions. It will generally come up with many solutions and grow to become more specific as more data is inserted. The best solutions are added to the systems while the ineffective solutions will be removed from the artificial systems all together. We can see that there are diverse forms of artificial intelligence systems that have been used in the field of medicine. Each system is useful for a different purpose in the branch of medical sciences. The existence of these varying systems ties back to the future possibilities concerning artificial intelligence. Just as students of varying thinking capabilities solve problems in their special way, machines can be created to operate data sets and variables to suit respective branches of medicine, showing that there is room for innovation and implementation of permanent residency of artificial intelligence in medicine. But just like humans, machines are not perfect.
Controversy Regarding Artificial Intelligence
Though there is vital evidence that concludes in initiating artificial intelligence in all aspects of medicine, Ramesh concludes that AI technology is still a sensitive topic and “it is the obligation of researchers active in this field to produce evidence that these techniques work on a practical level.” In other words, because AI technology is still a machine that produces outcomes based on a given set of data and variables, it cannot be trusted to make final decisions unlike a doctor that has studied, researched, and educated themselves in their particular fields. Thus, researchers must continue to process and revolutionize artificial intelligence to be proficient enough to be trusted to make final outcomes for a living patient. As a supporting factor in Ramesh’s conclusion, the journal “Artificial intelligence in medicine: current trends and future possibilities,” by authors Varun H Buch, Irfan Ahmed, and Mahiben Maruthappu published by the British Journal of General Practice in 2018 argues that “machines lack human qualities such as empathy and compassion, and therefore patients must perceive that consultations are being led by human doctors.” In other words, machines are robots and because they are not a living creature and the same biological race as humans, patients are very highly likely to mistrust a diagnostic coming from a non-living machine. A random comical example of why humans mistrust machines could also be because of the stigma revolving around AI taking our jobs, ultimately controlling businesses and taking over the world. In the article “Artificial intelligence in medicine,” Pavel Hamet, doctor, researcher, editor, administrator and teacher in Quebec, and Johanne Tremblay, doctor and Professor of Medicine at the University of Montreal, dives into a humans’ fear in machines. “The biggest apprehension we have is that AI will become so sophisticated that it will surpass human brain capabilities and eventually will take control over our lives.” Though it seems like a plot from a science fiction movie, the fear of machines with advanced artificial intelligence taking over the world is plausible. However, to tackle this issue researchers have to develop ethical standards in order for AI to function in a friendly manner, as to not inhibit protest from the public. As unlikely as that may be in present time, this terrifying belief that artificial intelligence will have a life of its own creates a shroud of mistrust among humans towards machines. But who can blame people for distrusting a non-living entity? In the article titled “AI revolution in medicine,” Harvard Staff Writer Alvin Powell mentions that AI comes with risks. “Poorly designed systems can misdiagnose.” When a misdiagnosis happens, the patient is being treated incorrectly and their condition may even worsen. For example, if a patient was diagnosed with cancer but improperly diagnosed a specific treatment, the treatment would be ineffective to the cancer and thus the cancer can grow to substantial lengths. Thus, it is extremely relevant to have doctors double check the statement of the AI and provide an expert diagnosis. Regardless of the mistrust, properly trained machines have proven themselves to excel at well-defined tasks, thus the future of artificial intelligence in the world of medicine is very bright.
The Future of AI in Medicine
As stated in the journal published by the British Journal of General Practice for the specific task of an AI to classify skin lesions, “the input is a digital photograph and the output is a simple binary classification: benign or malignant.” Previously mentioned in the article by Ramesh, the journal published by the British Journal of General Practice agrees that an AI’s ability to consume digital photos and produce outcomes that can classify skin lesions as benign or malignant is a popular advancement of artificial intelligence. Thus it appears that if digital imaging corresponds with a given field of medicine, artificial intelligence has a place in its community and can be implemented by the professionals of that field. Multiple articles have recognized the advancement of artificial intelligence and its ability to distinguish digital images to produce an educated doctoral diagnostic of the patient’s condition. Referring back to the scholarly journal, future artificial intelligence could “automatically prepare the most important risks and actions given the patient’s clinical records.” A patient’s clinical record can be simply or extremely complicated. Humans in the contemporary era can live up to 100 years of age. Imagine the considerable amounts of information that can be stored in that individual’s lifetime! By developing precise and powerful algorithms, AI could make doctor visits more efficient and prioritize patients that are (as processed prior to the appointment by an AI) in more critical conditions. A feature of AI that makes these developments possible is it’s ability “simultaneously monitor millions of inputs” AI is capable of having a significant role in preventing critical conditions in a patient. Advanced AI technology can introduce to patients a personalized medicine, specific to the patient and their history, which can ultimately lead to a healthier and lengthened lifespan. Artificial intelligence, brought up in the 20th century continues to grow throughout the 21st century and will continue to progress throughout the years to come. Though there are hesitations to the idea of implementing artificial intelligence in the healthcare industry, I believe that refined artificial intelligence that has been tested to its full potential will produce outcomes that are beyond the human imagination. That future may come sooner than we expect it to!
Citations:
Buch, Varun H, et al. “Artificial Intelligence in Medicine: Current Trends and Future Possibilities.” British Journal of General Practice, vol. 68, no. 668, 2018, pp. 143–144.
Hamet, Pavel, and Tremblay, Johanne. “Artificial Intelligence in Medicine.” Metabolism, Clinical and Experimental, vol. 69, 2017, pp. S36–S40.
Ramesh, A N, et al. “Artificial Intelligence in Medicine.” Annals of the Royal College of Surgeons of England., vol. 86, no. 5, pp. 334–338.
Powell, Alvin. “Risks and Benefits of an AI Revolution in Medicine.” Harvard Gazette, Harvard Gazette, 4 Dec. 2020, news.harvard.edu/gazette/story/2020/11/risks-and-benefits-of-an-ai-revolution-in-medicine/.