Prior to conducting my study, I had my doubts about artificial intelligence's applicability, particularly in the field of managing health information. I essentially imagined completely autonomous robots that were going to interfere with human existence when I thought of artificial intelligence. The reality of artificial intelligence, though, is very different from that assertion. Because of the advantages it can provide for numerous industries, including production and healthcare, it has emerged as the new frontier for researchers and developers. With the creation of LUCY to automate laboratory data administration and experiment planning, artificial intelligence (AI) for healthcare applications began to be developed as early as the 1980s. In the past few years, AI development has boomed in response to advances in computer hardware and software. AI is allowing for greater efficiency in labs and treatment of patients, reducing costs for healthcare administrators, and predicting possible outcomes of treatments or outbreaks. Tyagi (1) explains that AI transforms the nature of everything including macro-economic factors such as economy, employment, warfare, security and in this case, healthcare. It is imperative to understand the meaning of AI before venturing into its effect on the healthcare sector. In his report, Parloff Rodger (2), defines AI as a vast range of technologies that enable computers and robotic tech to solve problems that superficially resemble thinking.
The Initial Search and Obstacles
To begin with, my first interest in the subject was prompted by a report on CNN regarding how big data and artificial intelligence would revolutionize human lives, most especially in the healthcare industry. With the evolution of digital capacity more data is analyzed and stored in the digital space. The report mentioned that AI in medicine would assist in organize patient treatment plans and also provide physicians with literally all the data and all information that would enable them to give the right diagnosis.
Upon digging further, media outlets had received news about how Open AI had defeated one of the best players in DOTA2. (Albanesius 6). This sparked my interest on how AI achieved this. I initially started my search into artificial intelligence with a generic thought of “how artificial intelligence has benefited human society.” I had some inkling that the initial search would result in many sources to work with, but I was not prepared for just how large the results were. I decided to tackle this problem by selecting three key fields in which AI has had the greatest effect: automotive, healthcare and manufacturing. For instance, in the automotive sector, I read that Toyota Company is assembling a car that not only understands driving habits but also the kind of roads that the car can use (Krishna 1). Similarly, in the petroleum sector, GE (General Electric) has used AI in gathering and reassessing historic data in a more accurate way. The technology entails a deep diving data power that reassembles a whole world of information which significantly aids in exploration (Matthew, 2017). I continued my research focusing on AI in those three fields. I asked myself “how is software X helping field X” or “robotics isn’t really AI and if so, how can AI benefit them. This is where I hit another obstacle because many of my sources were either advertising AI or did not provide enough data to back up their claims. I decided to either narrow my topic to one field; healthcare. I had obtained 4 sources that could help me overcome the research obstacle. AI in medicine by Hamet, P & Trembay assisted me to narrow my research down to AI in the medical and healthcare fields.
Main Search
After choosing the topic, I continued with my research on AI, but this time with a narrow focus on how AI benefits medicine and health care environments. I used keywords such as Artificial intelligence, medical, medicine, healthcare, laboratory, lab automation, and robotics (Natarajan et al. 45). The sources that I was reading through after using the keywords allowed me to start a daily journal with what I learned. Alongside my journal, I used an I-Chart to maintain my focus on the topic. Some of my findings include:
Types and Benefits of AI.
There are two separate categories: virtual and physical
Virtual: software based AI
Benefits
Generating data from lab samples
The chances of errors are minimal due its great precision
Provide accurate results to doctors
Automates scheduling based on patterns at hospitals
Physical
Robots help in carrying out surgeries
Robots treat patients with mental health issues
It can help patients suffering from depression by keeping their brains active
Drawbacks of AI.
Over-relying on it may lead to technical deficiencies and hampers creativity. It can lead to false hopes
It reduces job openings and therefore, it replaces humans in many fields.
From my research, I came up with a number of conclusions. Artificial intelligence in healthcare has the prospects of becoming the backbone of every activity carried out in the healthcare sector. Firstly, it helps in mining medical records. This is the most obvious application of AI in a healthcare set up (Herper 3). Mining entails collecting, storage, normalizing and tracing of existing and past medical records. During my search, I came across the ‘Google Deepmind Heaalth project”. It is basically an application that is used in mining data from medical records in a bid to make health services to be more efficient. Another important function of AI is to design treatment plans. IBM Watson recently launched a special program for oncologists and while reading his work, I was able to deduce that AI creations provide clinicians with evidence-based treatment options. AI is essential in analyzing the framework of structured and un-structured data in clinical reports that are vital in selecting treatment pathways. This is done by combining the attributes in the patient’s and relating them with external research, clinical expertise and data. Consequently, the program is able to identify potential treatment plans that best suits the patient (Herper 1).
Among other notable functions of AI in healthcare includes assisting in repetitive jobs, accelerating performance of online consultations, health assistance, medication management and drug creation. The artificial intelligence interface uses machines to support patients who suffer from chronic conditions. Also, it helps in customized monitoring and follow up care. Moreover, AI has an impact on precision medicine especially in the fields of genetics and genomics (Luxton 34). Deep genomics identifies patterns that come in huge data sets entailing genetic information and medical records. Such computational technologies can help in altering their DNA either naturally or therapeutically.
According to Robert Dostie (a senior director at Carestream Health), health care companies are launching healthcare products and applications that use machine learning algorithms and predictive analytics that help in reducing recovery times and providing virtual assistance in diagnosing ailments. Technology has enhanced the discovery of diabetes management tools that use interface to connect physicians and patients. The AI technology used is known as an insulin-dose decoder and it is compatible with a number of insulin pens. It captures injection data and it has an app maintained using a smartphone or a computer (Sullivan 1).
During my research, I came across a fascinating technology known as 3D printing. This AI technology enables practitioners in carrying out CT scans, MRI and development of a deeper insight into a patient’s anatomy (Franco 14). Other 3D printing models stimulate organs with differing anatomical difference; a function that surgeons to evaluate the patient’s condition and determine the best way to perform a procedure. General Electric healthcare division is charged with development of artificial intelligence products for medicine. It manufactures imaging devices to be used by radiologists in focusing and in DNA altering.
C.J. Rawlings (2017) explores AI in molecular biology. A number of software have been put to in the general field of molecular biology. The software ensure protein structure prediction and in molecular genetics. In molecular genetics, they create integrated knowledge base that simulates biochemical metabolism and identify nucleic acid sequences (Rawlings, 2017).
Other aspects of my research led me to Sangwook Kim’s study on neural networks. He mentions the importance of developing advanced human-robotic interaction system in order to understand how humans think, perceive and act. Consequently, this knowledge will be helpful in human mental development and psychological interaction. Robots can provide artificial cognitive agents that aid in such neurological interactions (Sangwook K., Zhibin, Y. & Lee, M 1). On the same note, Le Cun & Brockman (12), argue that artificial intelligence is a field that most tech companies are racing towards in a bid to create machines that think like people. Take Facebook for instance, the company is a prime mover behind deep neural networks (LeCun & Brockman 1).
In addition, Radick (23) discusses the importance of large scale deployment of AI. Therefore, artificial intelligence in healthcare is important in two ways. First, it performs activities that humans do but with more consistency and quality. Secondly, it provides new solutions that humans cannot decode due massive amount of data involved in the analysis. For instance, AI-simulator is used in surgical training; an approach that would be much harder if the process was conducted using a trainer. I located an interesting excerpt from Pavel and Tremblay’s article titled artificial intelligence in medicine. They mention that AI combines adaptive evolutionary algorithms and advanced clustering methods I the determination of DNA variants and other nucleotide polymorphisms (12). Furthermore, it is used in the creation of softbots, a psychotherapeutic avatar that control pain in children with cancer.
Another resource was Ben Dickson’s book, how AI is revolutionizing the mHealth industry. For instance, he mentions the usefulness of mHealth apps as a source of information for self-care provision. An example is Ada, a mobile app that uses conversational chat-bots to aid patients in identifying symptoms of certain diseases. It identifies patient related data such as medical history and risk related factors. It also helps in mitigating risks and damage of self-diagnosis (Dickson 2). To summarize, it is important to note how health information technology improves aspects of care such as efficiency, quality and medical care delivery. More so, it helps in managing administrative, financial and multi-functional health records. Essentially, it automates labor intensive and inefficient processes (Isern & Moreno 56).
My Final Thoughts on AI
We should focus on balance between human to AI interaction
I approve the use of AI due to the aforementioned benefits.
Data needs to be open sourced, but only raw data (no names, or personal details about patients or the samples used to get data).
References
Dickson, Ben. How Artificial Intelligence is Revolutionizing the Mhealth Industry. UK National Health Service. London, 2016
Franco, G.. Artificial intelligence in Microbiology for faster actionable results. MLO: Medical Laboratory Observer, 49(8), 54-56. 2017
Hamet, Pavel & Tremblay, Johanne. Artificial Intelligence in Medicine. Metabolism Clinical and Experimental 69, Elsevier, 2016
Herper, Matthew. GE Partners With Harvard Hospitals To Harness Artificial Intelligence In Medicine. Forbes.Com, 1, 2017.
Isern, David & Moreno, Antonio. A Systematic Literature Review of Agents Applied in Healthcare. Springer Science. New York, 2015.
Krishna, Vishak. Artificial Intelligence: Why Toyota and other car makers want this tech in by 2030. Retrieved from: contify.com, 2017.
Luxton, David D. Artificial Intelligence in Behavioral and Mental Health Care. , 2016. Internet resource.
Matthew, Stevens. Mining’s tech allies could disrupt sector. The Australian Financial Review: Melbourne. Proquest LLC, 2017.
Natarajan, Prashant, John C. Frenzel, and Detlev H. Smaltz. Demystifying Big Data and Machine Learning for Healthcare. , 2017. Print.
Parloff, Roger. The Deep Learning Revolution. Fortune.cpm. Time Inc., 2016.
Radick, Lea. Artificial Intelligence in Healthcare: The Current Compelling Wave of Interest. Healthcare Executive, 2017. 21-29
Rawlings,, J. P. Fox., E. A. Thompson, A, & B. Robson. Artificial Intelligence in Molecular Biology: A Review and Assessment [and Discussion]. Philosophical Transactions: Biological Sciences, (1310), 353, 2017
Sangwook, Khan., Zhibin, Yu. & Minho, Len. Understanding human intention by connecting perception and action learning in artificial agents. Kyungpook National University, 2017.
Sullivan, Tom. Artificial Intelligence in Healthcare. Journal of AHIMA. 2017.
Tyagi, Amit. Artificial Intelligence: Boon or Bane, 2017.
Yann, Lecun & Brockman, Greg. Artificial Intelligence is Smarter when it’s open Source. Brockman CTO. Facebook. Openal, 2017.
