Artificial Intelligence in Medicine
AI – base on – accurate decision-making
แบ่ง subdivided – strong AI , – conscious and intentionality
weak AI – lake both , program to perform specific task only
. With AI currently making rapid progress all domains, including those of healthcare, physicians face possible competitors – or worse, with claims that doctors may become obsolete. Various types of AI programmes are already available as consultants to the physician, and these help in medical diagnostics and treatment. At the time of writing, extant programmes constitute weak AI. This paper will explore the development of AI and robotics in medicine, and will refer to Star Trek’s “Emergency Medical Hologram,” who is portrayed as a strong AI programme. This paper will also briefly explore the issues pertaining to AI in the medical field and will show that weak AI should not only suffice in the demesne of healthcare, but may actually be more desirable than strong AI.
Introduction
Artificial Intelligence (AI) is a relatively new field in science and engineering. Work in this domain commenced in earnest soon after the Second World War, and the term Artificial Intelligence was coined in 1956 (Russell and Norvig 18). AI currently encompasses a wide array of subfields, ranging from the general, such as learning and perception, to the specific, such as chess, proving mathematical theorems, writing poetry and diagnosing diseases.
AI is based on an accurate decision-making process which can be carried out independently by a machine. The study of mechanical reasoning has a long history. Chinese, Indian and Greek philosophers all developed structured methods of formal deduction in the first millennium BCE. Majorcan philosopher Ramon LLul (1232-1315) developed several machines devoted to the production of knowledge by logical means (Bonner 60). Llul described his machines as mechanical entities that could combine basic and undeniable truths by simple logical operations so as to collect all the possible knowledge (Bonner 60). Leibniz envisioned a universal language of reasoning so that ‘there would be no more need of disputation between two philosophers than between two accountants. For it would suffice to take their pencils in hand, down to their slates and says to each other: Let us calculate’ (McCorduck 41).
The earliest research into thinking machines was inspired by a confluence of ideas that became prevalent in the late 1930s, 40s and early 50s. Research in neurology had shown that the brain was an electrical network of neurons that fired in all-or-nothing pulses. Norbert Weiner’s theory of cybernetics described control and stability in electrical networks, while Claude Shannon’s information theory described digital signals. Furthermore, Alan Turing’s theory of computation conjectured that any form of computation could be described digitally. The close relationship between these ideas suggested that it might be possible to construct an electronic decision-making brain (McCorduck 55).
The first modern computers were massive code breaking machines used during the Second World War for the purposes of cryptography. They were based on the theoretical foundation laid by Turing and further developed by John Von Neumann (Mc Corduck 77). In 1950, Alan Turing published a landmark paper in which he speculated about the possibility of creating machines that think, and introduced the concept of the Turing test (Turing 433). This gedankenexperiment allowed Turing to argue convincingly that a thinking machine was at least plausible, and the paper answered all of the most common objections to the proposition (Norvig and Russell 948).
Fundamental to the ambitions of artificial intelligence (AI) as a typical example of the step from logical foundations to technology is a thesis of computational sufficiency (MunKim 77). MunKim further contends that there is another version on which many AI researchers rely. It is the so called AI thesis which says: ‘As the intelligence of the machines evolves, its underlying mechanisms will gradually converge to the mechanisms underlying human intelligence (MunKim 71). From the beginning, AI researchers did not shy from making predictions of the anticipated coming successes in the development of strong AI. The following statement by Herbert Simon in 1957 is often quoted:
It is not my aim to surprise or shock you – but the simplest way I can summarize is to say that there are now in the world machines that think, that learn and that create. Moreover, their ability to do these things is going to increase rapidly until – in a visible future – the range of problems they can handle will be co-extensive with the range to which the human mind has been applied (Simon 198).
Simon made concrete predictions, that within ten years, AI would have flourished. These predictions came true within forty years not ten. Russell and Norvig purport that Simon’s overconfidence was due to the promising performance of simple, early AI systems (Russell and Norvig 21).
However, with AI currently making vast and uninterrupted progress in the domain of medical health and elsewhere, physicians are now facing possible peers amidst claims that doctors may become obsolete. This paper aims to explore the juxtaposition of medicine and AI and the impact AI has and will continue to have on medicine. It will also seek to analyze the implications this will have on medical doctors and healthcare.
Artificial Intelligence in Medical Diagnosis and Interventions
The steady expansion of medical knowledge has made it progressively more difficult for the physician to remain abreast of medicine outside of a chosen specialty. The steady expansion of medical knowledge has made it progressively more difficult for the physician to remain abreast of medicine outside of a chosen specialty. Campbell alleges that this happens because knowledge is comprised of a ‘continuous texture of narrow specialties, […] a redundant piling up of highly similar specialties leaving interdisciplinary gaps’ (Campbell 328).
By the early 1970s it became clear that conventional tools such as flow charts, pattern matching, and Bayes’ theorem were unable to deal with more complex clinical problems (Gorry 50). Investigators thus began to study the expert physician in order to obtain detailed insights into the nature of clinical problem solving (Kassirer and Gorry 250; Swanson et al, 160). The results derived from such studies have subsequently formed the basis for computational models of cognitive phenomena, and these models have further been converted into so-called artificial intelligence programmes (Weiss et al, 150).
Many of the early efforts to apply artificial intelligence methods to real problems, including medical reasoning, have primarily used rule-based systems (Duda and Shortcliffe 263). Such programmes are typically easy to create, since their knowledge is catalogued in the form of ‘if-then’ rules in chains of deduction that reach a conclusion. However, most serious clinical problems are broad and complex such that straightforward attempts to chain together larger sets of rules often encounter major difficulties. Problems arise principally from the fact that rule based programmes do not embody a model of disease or clinical reasoning (Duda and Shortcliffe 263).
Given the difficulties encountered with rule based systems, more recent efforts to use AI in medicine have focused on programs organized around models of disease. Efforts to develop such programmes have led to substantial progress in the conversion of various models into promising experimental programmes. These improved representations of clinical knowledge and sophisticated problem solving strategies have significantly advanced the field of artificial intelligence in medicine.
Any programme designed to serve as a consultant to the physician must contain certain basic features. It must have a store of medical knowledge expressed as description of possible diseases. Depending on the breadth of the clinical domain, the number of hypothesis in the database can range from a few to many thousands. In the simplest conceivable representation of such knowledge, each disease hypothesis identifies all the features that can occur in a particular disorder. In addition, the programme must be able to match what is known about the patient within its store of information (Dilsizian and Siegel 442).
Medical diagnostic programmes similar to the Virtual Space Station, could serve as useful tools in today’s overburdened health care system. For instance, IBM’s Watson – best known for its remarkable performance on Jeopardy! – is now being used in healthcare applications. It provides many unique and transformative possibilities to resolve challenges associated with medical diagnosis and treatment. The Watson hardware costs approximately three million dollars. It can process 500 gigabytes of data per second, the equivalent of one million books (Pearson). Such medical applications may help physicians navigate through a complex set of patient symptoms, laboratory data and imaging results to come up with a set of ‘most likely’ clinical diagnoses and treatment options. Software of this nature may ultimately improve patient outcomes and reduce health care costs (Dilsizian and Siegel 444). IBM initially utilized the American College of Physicians Medical Knowledge Self-Assessment Study Guide, and subsequently improved on its performance by adding textbooks, such as the Merck Manual of Diagnosis and Therapy, and additional medical journals included in the original database. The developers then further improved on the performance for medical applications by fine-tuning the weighting associated with the various algorithms utilized by the application for these medical domain questions. The team also created a demonstration in which a patient’s presenting symptoms are input into the software and a series of progressive questions are posed by a health care worker to personalize the diagnostic and therapeutic recommendations made by the software (Pearson). Two important features of the prototype software were the ability to provide multiple possible diagnoses and treatment options with relative confidence levels, as well as the ability to trace the information utilized to make a recommendation (Dilsizian and Siegel 444).
AI is also being used in the field of cardiology, including the determination of the most appropriate type of imaging study for a specific set of symptoms (Fornell 35). For example, the Imaging in FOCUS (Formation of Optimal Cardiovascular Utilization Strategies) quality improvement, an initiative of the American College of Cardiology, was recently introduced so as to channel the utilisation of diagnostic imaging through the arrangement of AI that tracks appropriate use criteria (Saifi et al, 823). Fifty-five participating sites voluntariy completed the radionuclide imaging performance improvement module. Results showed that the proportion of inappropriate cases decreased from 10% to 5%. These preliminary data suggest that the use of self-directed, quality improvement software and an interactive community may permit physicians to significantly decrease the proportion of tests not meeting appropriate use criteria (Saifi et al, 825).
After images are acquired, additional AI tools may help physicians provide accurate interpretation of cardiac imaging studies (De Puey et al, 1165). Artificial neural networks are thus excellent example of ways in which current AI systems that emulate human decision-making can be successfully applied in cardiac imaging (Itchhaporia et al, 517). These networks have been utilized in the diagnosis and treatment of coronary artery disease and myocardial infarction, the interpretation of electrocardiographic studies and the detection of cardiac arrhythmias, such as ventricular fibrillation (Clayton et al, 219).
These expert systems are costly, difficult to develop and maintain, and require a perfect match between input data and existing rule forms. Software such as that utilized by Watson, on the other hand, uses natural language processing and a variety of search techniques to create hypotheses, making it more flexible, scalable, easy to maintain, and cost effective. This new approach makes it much easier to keep up with ever-changing information in imaging, medicine and surgery.
In a future clinical image interpretation scenario utilizing AI technology, a requested study would first be evaluated for appropriateness based on the patient history and previous examination. The examination would, if deemed appropriate, then be ‘protocoled’ with regard to the way in which it should be performed. For example, a magnetic resonance scan would be fine-tuned as to imaging sequences and amount and type of radiopharmaceutical and/or contrast material to optimize efficacy, safety and efficiency of the investigation (Clayton et al, 221).
Technology is also revolutionizing the medical field through the creation of robotic devices and complex imaging. Robotic surgery, computer-assisted surgery and robotically-assisted surgery are terms for technological developments that use a robotic system to aid in surgical procedures. Robotically-assisted surgery was developed to overcome the limitations of pre-existing minimally-invasive surgical procedures and to enhance the capabilities of surgeons performing open surgery.
The first documented use of a robot-assisted surgical procedure was in 1985 when the PUMA 560 robotic surgical arm was used in a delicate neurosurgical biopsy, an open and therefore non-laparscopic procedure (Kwoh et al, 155). The system allowed successful robotic surgery and the potential for greater precision when used in minimally invasive surgeries. This led to the first laparoscopic procedure in 1987, involving a robotic system to carry out a cholecystectomy (gall bladder removal) (Jones and Jones 120). The following year, the same PUMA system was used to perform a robotic surgical transurethral prostate resection. In 1990, the AESOP system produced by Computer Motion became the first system approved by the Food and Drug Administration (FDA) for endoscopic procedures (Jones and Jones 123).
In the case of robotically-assisted minimally-invasive surgery, instead of directly moving the instruments, the surgeon uses one of two methods to control the instruments; either a direct tele-manipulator or through computer control. A tele-manipulator is a remote manipulator that allows the surgeon to perform the normal movements associated with the surgery whilst the robotic arms mimic the same movements to operate on the patient. In computer-controlled systems the surgeon uses a computer to control the robotic arm and its end-effectors, though these systems can still use tele-manipulators for their input. One advantage of using the computerized method is that the surgeon does not have to be present, with the possibility for remote surgery. One minimally invasive option is the da Vinci surgery system which has brought minimally invasive surgery to more than three million patients worldwide (Samadi). In 2000, the da Vinci system broke new ground by becoming the first robotic surgery system approved by the FDA for general laparoscopic surgery (Samadi). This was the first time the FDA approved an encompassing system of surgical instruments and camera-scopic utensils. The da Vinci system’s three dimensional magnification screen allows the surgeon to view the operative area with the clarity of high resolution. The one centimeter diameter surgical arms represent a significant advancement in robotic surgery from the early-large-armed systems such as the PUMA 560 (Samadi). This advance allows for less contact between exposed interior tissue and the surgical device, greatly reducing the risk of infection. The ‘Endo-Wrist’ features of the operating arms precisely replicate the skilled movements of the surgeon at the controls, improving accuracy in small operating spaces (Samadi).
Robotic surgery is at the cutting edge of precision and miniaturization in the realm of surgery, thus the possible applications are as extensive as the uses of minimally invasive surgery. Robotic surgery has already become a successful option in neurological, gynecological, cardiothoracic and numerous general surgical procedures (Samadi).
It is reasonable to assume that the current advantages of robotic surgery systems will be expanded with the next generation of medical robotics. Removing human contact during surgery may be taken to the next level with robotic surgery systems capable of functioning at greater distances between surgeons’ control consoles and the patient side table robotics. This would allow robotic surgery to be conducted with patients in a nearby ‘clean room,’ reducing the risk of intraoperative infection. Major strides are also being made in creating robotic surgery systems which are more capable of replicating the tactile feel and sensation which a surgeon experiences during more invasive traditional procedures, allowing the operator the precision and advantages of minimally invasive procedures without losing the sensory information that is so helpful in making judgment calls during robotic surgery.
Weak AI and Strong AI
There are two major ways to think about the current utilization and power of artificial intelligence. The weak AI hypothesis states that a machine running a programme is at most only capable of simulating real human behaviour and consciousness. Artificial intelligence such as that being used in medical diagnosis and certain interventions are examples of weak AI because they are focused on one narrow task. Weak AI justifies the claims that a running AI program is at most a simulation of a cognitive process but is not itself a cognitive process. Strong AI, on the other hand, purports that an (as yet to be written) programme running on an (as yet to be designed) machine is actually a mind – that there is no essential difference between a piece of software exactly emulating the actions of the brain, and the actions of a human being, including their understanding and consciousness.
The debate between the two opposing forces of weak AI and strong AI is intense, with scientists and philosophers of both camps foregrounding claims to support their hypotheses. John Searle’s Chinese Room Argument (CRA) is a celebrated thought experiment designed to refute the hypothesis, that “the appropriately programmed computer really is a mind” (Searle 420) First formalized in 1980 in a target article in Behavioural and Brain Sciences, it was designed to show the futility of the search for Strong AI. Searle envisages a situation in which he is hidden in a room and is presented questions in Chinese written on an input card, posted in to his room by unseen enquirers. Searle knows no Chinese, indeed, he is quite unaware of the enterprise in which he is engaged and is ignorant of the fact that the strange marks on the cards represent questions framed in Chinese (Searle 451). He consults a manual telling him (in English) precisely what equally strange marks to write on an ‘output’ card, which he posts back to the outside world. By virtue of the ‘machine intelligence’ embodied in the manual, these marks on the output card constitute an answer to any input question. To a Chinese speaker external to the room, by virtue of its question answering ability, the system passes the Turing test for machine intelligence (Turing), yet the system implemented by Searle-in-the-room is entirely without understanding simply because Searle does not understand Chinese (Searle 455).
Searle concludes that an AI program could give the impression of intelligence to an external observer, but completely lack understanding. John Searle and Ned Block contend that even if a system duplicates our behaviour, it might be missing important ‘internal’ aspects of mentality: consciousness, understanding and intentionality (Searle; Block). Thus a weak AI is only capable of simulating real human behaviour and consciousness, without intentionality. Edmund Husserl called this ‘the principle theme of phenomenology’ (Husserl 146). This is also supported by Searle’s famous Chinese argument which holds that a program cannot give a ‘mind,’ ‘understanding,’ or ‘consciousness, regardless of how intelligently the combination behaves’ (Searle). Searle further purports that causal features of the brain are necessary and crucial for intentionality.
“Could a machine think?” […] only very special kinds of machines, namely brains and machines with internal causal powers equivalent to those of brains. And that is why strong AI has little to tell us about thinking, since it is not about machines but about programs, and no program by itself is sufficient for thinking (Searle 417).
Searle upholds ‘biological naturalism.’ He explains that human mental phenomena such as consciousness and understanding require a specific biological process that is found in brains of human beings, an important combination of physical and chemical properties. Thus a man-made intelligent system that acted exactly like a human mind might still not be conscious (Searle 264). It is therefore Searle’s contention that strong AI may share our behavioural and functional equivalence, without being a conscious system, since consciousness requires not only a computational organization but also a specific and unknown way in which functional organization is implement in the biology of the organism. This is because Searle believes that the human brain is able to conjure ‘perceptual aboutness’ (Natsoulas 76) and that:
the brain’s causal capacity to produce intentionality cannot consist in its instantiating a computer program, since for any program you like it is possible for something to instantiate that program and still not have any mental states. Whatever it is that the brain does to produce intentionality, it cannot consist in instantiating a program since no program, by itself, is sufficient for intentionality (Searle 424).
Colin McGinn further argues that the famous Chinese Room gedankenexperiment provides strong evidence that the debate revolves around whether a machine can be shown to actually be conscious as opposed to the theory that a machine can be conscious (McGinn 194). This type of AI is even more commonly depicted in popular culture, recently and famously in Ex Machina (Garland) and Transcendence (Pfister). Searle, however, doubts that true consciousness in an android could ever be possible with our present state of knowledge.
This is contrary to the tenets of strong AI – that computational states are functionally equivalent to mental states (Schank and Abelson; Mc Carthy). DeDompablo Cordio opines that strong AI itself seem to be rooted in large part in science fiction (52). This echoes a prototype of strong AI – the Emergency Medical Holographic program (EMH) created in the fictional Star Trek future. The EMH is portrayed as a sophisticated hologram developed in the early 2370s and used on most Federation starships in the late 24th century. It was designed to provide short-term advanced assistance during emergencies in sick bay to the extent of literally replacing a starship’s medical officer. The EMH Mark 1 was first activated in 2371, and was programmed with over five million possible treatments from the collective information of 2000 medical references and the experience of forty seven individual medical officers. The EMH was also supplemented with contingency programs and adaptive programs to learn while serving as a supplement of a normal medical staff in cases of emergency. It contained fifty million “gigaquads” of computer memory, which is “considerably more than most highly developed humanoid brains” (Bole “Lifesigns”).
While the EMH may seem preposterous, a technological ‘singularity’ may make it possible. The term ‘singularity’ was introduced by the science fiction writer Vernor Vinge in a 1983 opinion article. It was brought into wider circulation by Vinge’s influential 1993 article “The Coming Technological Singularity” (Vinge) and by the inventor and futurist Ray Kurzweil’s popular 2005 book The Singularity is Near (2005). The concept of Singularity was succinctly set out by the statistician I. J. Good in his 1965 article “Speculations Concerning the First Ultra-intelligent Machine.’’
Let an ultra-intelligent machine be defined as a machine that can far surpass all the intellectual activities of any man however clever. Since the design of machines is one of these intellectual activities, an ultra-intelligent machine could design even better machines; there would then unquestionably be an ‘intelligence explosion’ and the intelligence of man would be left far behind. Thus the first ultra-intelligent machine is the last invention that men need ever make (Good 31).
David Chalmers further argues that this intelligence explosion may be combined with another potentiality, the ‘speed explosion.’ His contention lies in Gordon Moore’s observation that a computer processing speed doubles at regular intervals (Moore 115). The singularity proposed by Chalmers depends on an uncritical acceptance of the ‘assumption that there is such a thing as intelligence and that it can be measured’ (Chalmers 16). However, there are many different ways of evaluating cognitive agents, no one of which deserves the canonical status of ‘intelligence.’ Furthermore, if there is a canonical notion of intelligence that applies within the human sphere, how this can be extended to arbitrary non-human system, including artificial systems is far from clear.
McDermott takes this a step further when he claims that personal identity is mostly a matter of social convention. McDermott claims that to dwell too much on this philosophy of singularity would lead the human race to miss other possible and more important events (McDermott 5). A similar kind of resistance which Mc Dermott reveals was also faced by the EMH in Star Trek. In an earlier Star Trek episode, when Mr Spock lamented to Dr Leonard Mc Coy that computer programming had had not yet advanced far enough to replace a ship’s doctor, Mc Coy retorted that he would rather resign than be replaced (Lucas, “The Ultimate Computer”).
Chalmers believes there are obstacles to the Singularity, with the most serious opposing force being what he calls a ‘motivational defeater’ (Chalmers 21). Chalmers purports that it is entirely possible that most humans will be disinclined to create AI. It is entirely possible that there will be active prevention of the development of strong AI, although Chalmers contends this could not be prevented indefinitely (Chalmers 22). Chalmers reiterates that in any case, it is not a question of ‘if’ but ‘when.’ Humanity therefore needs to think hard about the form it should take. The most obvious question is arguably what sort of values need to be instilled in a non-human-based AI (Chalmers 32). Beyond the Asimovian maxims of human survival, well-being and obedience to human command, strong AI should also arguably value scientific progress, peace and justice among other ideals. This might proceed either by a higher-ordering valuing of the fulfillment of human values or by a first – order valuing of the phenomena themselves. This hierarchy of values largely assumes that intelligence and values are independent of each other. David Hume advocated a view on which value is independent of rationality: a system might be intelligent and rational, while still having arbitrary values. By contrast Immanuel Kant advocated a view on which values are not independent of rationality: some values are more rational than others. Thus if intelligence correlates with rationality, an intelligence explosion may be accompanied by a morality explosion (Chalmers 46).
Some might argue that the ideal would be to have a perfect replica of Homo sapiens, with identical physical and psychological attributes instantiated in a strong AI system. However, the pressing point here is whether there will be an eventual desire or even necessity for strong AI to replace medical doctors. If an AI is programmed to possess two fundamental characteristics, flawless computation and indefatigability so as to reach an accurate diagnosis treatment, AI would have reached its ultimate medical goal. The same characteristics are noted in Star Trek’s EMH – a holographic simulation of a human doctor which was present on the starship USS Voyager and which ran almost continuously for seven years.
These attributes all point to the thesis that ultimately, AI will function independently and will constitute a major part of the work force. Thus, this strongly suggests a major transition in the medical health field, where doctors, nurses and the rest of the multidisciplinary team will be replaced with stronger intelligences. One obvious counterargument is that a strong AI will be emotionless, lacking compassion – a paramount ingredient to care. Emotion is crucial and is inextricably tied to everything we say and do. Instantiating compassion modules as software-hardware components of such AI systems would make such systems far more acceptable. Of interest to AI research and psychology, is the idea of simulating emotion in computers. While psychologists have known for some time that there are a good deal of physical correlates to emotion – voice changes, blushing, pupil dilation etc – reproducing them continues to prove difficult. Emotion provides us with a motivation and drive, and with a set of personal preferences, a uniqueness that would desirable in a sophisticated AI (Maciamo).
The sociable Machines Project has developed an expressive anthropomorphic robot called Kismet that engages people in natural and expressive face-to-face interaction. Inspired by infant social development, psychology, ethology and evolution, this work integrates theories and concepts from these diverse viewpoints in order to enable Kismet to enter into a natural and intuitive social interaction with a human caregivers, and to learn from them (Breazeal 10). In order for Kismet to do so, it has the ability to perceive a variety of natural social cues from visual and auditory channels, and to deliver social signals to the human caregiver through gaze direction, facial expression, body posture and vocal babbles. The robot has been designed to support several social cues and skills that could ultimately play an important role in socially situated learning with a human instructor (Breazeal 10) Work done on internal emotion, most notably MIT’s Kismet’s two emotions (internal and external) are combined into a robot that displays emotion (Breazeal 9).
Sociable humanoid robots pose a dramatic and intriguing shift in the way one considers the control of autonomous robots. Traditionally, autonomous robots are designed to operate as independently and as remotely as possible from human. However, a new range of application domains (domestic, entertainment, health care) are driving the development of robots that can interact and co-operate with people, and thus play an active part in their lives (Brezeal 10). With this in view, nothing stands in the way of an AI medical program replacing the traditional human medical doctor.
Conclusion
James P. Hogan’s 1979 novel The Two Faces of Tomorrow is an explicit description of Singularity (Hogan). An artificial intelligence system solves an excavation problem on the moon in a brilliant and novel way but nearly kills a work crew in the process. Realizing that systems are becoming too sophisticated and complex to predict or manage, a scientific team sets out to teach a sophisticated computer network how to think more humanly. The story documents the rise of self-awareness in the computer system, the humans’ loss of control and failed attempts to shut down the experiment as the computer desperately defends itself, and the computer intelligence reaching maturity. This novel illustrates the successes and pitfalls which the real world might eventually face when technology progresses from the current developments of weak AI to potential developments of strong AI.
Chalmers purports that once a strong AI begins to function independently, the only viable option for human beings is ‘integration,’ where human beings become ‘superintelligent systems’ themselves (Chalmers 33). Chalmers disputes that:
In the long run, if we are to match the speed and capacity of non-biological systems, we will probably have to dispense with our biological core entirely. This might happen through a gradual process through which parts of our brain are replaced over time; or it happens through a process. Either way, the result is likely to be an enhanced non-biological system, most likely a computational system (Chalmers 33).
Driving his point home, Chalmers also claims that it is entirely possible that there will be active prevention of the development of strong AI, although he says this could not be prevented indefinitely (Chalmers 22). He alleges that human beings need to learn how to integrate with strong AIs as he sustains that an ‘intelligence explosion’ will happen in the coming decades, changing the conventional ways individual are accustomed to (Chalmers 10).
However weak AI would actually suffice for medical purposes. Such systems do not need to have intentionality and self-awareness to function well. The present AI programmes available on the market come with a computational sufficiency and serve as useful tools in today’s overburdened health care system. Medical aided programmes are providing correct diagnoses within minutes if not seconds, while robotic surgery is making huge advances in the field. This suggests that a strong AI might not even be necessary. If these systems are attainable now, the next obvious question is whether strong AI is truly necessary in the medical fields. If professionals, especially those in the medical field are served well with a weak AI, how important is it to create a strong AI for medical purposes?
Furthermore a strong AI might not necessarily be better in this demesne. A weak AI would controllable due to lack of intentionality and independence, a mind of its own. The lack of consciousness and intentionality in a weak AI make it what it simply is – a programmable machine. On the other hand, a strong AI will be able to function independently, possessing the attributes of free will and the moral conscience to choose right from wrong. The dilemma lies here: while a strong AI might be appealing for tomorrow’s world because it fulfills all the requirements being sought in the Singularity, a weak AI is what human beings might actually need. Homo sapiens has reigned supreme in the animal kingdom since the beginning of his existence – relegating any portion of his power to a strong AI might not be on his agenda for many years to come, making him think twice about the creation of such machines.
References
Cinematography
Ex Machina. Dir. Alex Garland. Universal Pictures. 2015.
“Lifesigns.” Dir. Cliff Bole. Star Trek: Voyager. USA: Paramount Pictures. 1996 February.
“The Ultimate Computer.” Dir. John M. Lucas. Star Trek: The Original Episode. USA: Paramount Pictures. 2008 February.
Transcendence. Dir. Wally Pfister. Warner Bros. Pictures. 2014.
Primary Texts
Block, Ned. “Psychologism and behaviourism.” Philosophical Review. 90 (1981): 5-43.
Bonner, Anthony. The Art and Logic of Ramon Llul: A User’s Guide. US: Brill, 2007.
Breazeal, Cynthia. “Emotion and Sociable Humanoid Robots.” Elsevier Science. (2002): 1-42.
Campbell Donald T., ‘Ethnocentrism of Disciplines and the Fish Scale Model of Omniscience’, in Interdisciplinary Relationships in the Social Sciences, ed. by Muzafer Sherif and Carolyn W. Sherif (Chicago: Aldine Publishing Co. 1969
Chalmers, David J. “The Singularity: A Philosophical Analysis.” Journal of Consciousness Studies. 17.7 (2010): 7-65.
Clayton, R; Murray, A; Campbell, RWF. “Recognition of Ventricular Fibrillation using Neural Networks.” Medical and Biology Engineering and Computing. 32(1994): 217-220.
DeDompablo Cordeo, Mariano. “Searle’s Chinese Room Argument and its Replies: A Constructive Re-Warming and the Future of Artificial Intelligence.” Indiana Undergraduate Journal of Cognitive Sciences. 3(2008): 51-57.
De Peuy, G.E; Garcia, E.V; Ezquerna, N.F. “Three-dimensional techniques and artificial intelligence in thallium – 201 cardiac imaging.” American Journal of Roentgenology. 152(1989): 1161-1168.
Dilsizian, Steven E and Siegel, Eliot L. “Artificial Intelligence in Medicine and Cardiac Imaging: Harnessing Big Data and Advanced Computing to Provide Personalized Medical Diagnosis and Treatment.” Nuclear Cardiology. 16 (2014): 441-447.
Duda, R.O and Shortlife, E.H. “Expert Systems Research.” Science. 220 (1983): 261-268.
Fornell, D. “An Introduction to Clinical Decision Support for Cardiology.” Diagnosis and Intervention Cardiology. 2003. Available at: www.dicardiology.com/article/introduction-clinical-decision-support-cardiology. Accessed 21 February, 2016.
Good, Irving J. “Speculations Concerning the First Ultraintelligent Machine.” Advances In Computers. 6 (1965): 31-38.
Gorry, G.A. “Computer-assisted Clinical Decision Making.” Methods of Information in Medicine. 12 (1973): 45-51.
Hogan, James P. The Two Faces of Tomorrow. Milwaukie: Dark Horse Manga. 1979.
Itchhaporia, D; Show, P.B; Almassy, R.J; Oetgen, W.J. “Artificial Neural Networks: Current Status in Cardiovascular Medicine.” Journal American College of Cardiology. 28 (1996):515-521.
Jones, S.B and Jones, D.B. “Surgical Aspects and Future Developments in Laparoscopy.” Anesthiology of Clinical North America. 19 (2001): 107-124.
Kassirer, J.P and Gorr, G.A. “Clinical Problem Solving: A Behavioural Analysis.” Annals of Internal Medicine. 89 (1978): 245-255.
Kurzweil, Ray. The Singularity is Near. USA: Penguin Group. 2005.
Kwoh, Y.S; Hou, J; Jonckheere, E.A; Hayall, S. “A Robot with Improved Absolute Positioning Accuracy for CT Guided Stereotactic Brain Surgery.” IEEE Transactions on Biomedical Engineering. 35.2 (1988): 153-161.
Maciamo, Hay. “Could a machine or an AI ever feel human-like emotions?” Humanity Media. 2014. Available at www.hplusmagazine.com Accessed 29 February 2016.
McCarthy, John. “Ascribing Mental Qualities to Machines.” Introduction (1979). Reprinted in Formalizing Common Sense: Papers. New Jersey: Ablex Publishing Corporation. 1990.
McCurdock, Pamela. Machines Who Think. (2nd ed). USA: A.K. Peters Ltd. 2004.
McDermott, Drew. “Response to ‘The Singularity: A Philosophical Analysis.” Available at www.cs.yale-edu./homes/dvm/papers/chalmers-singularity-response Accessed 25 February, 2016.
McGinn, Colin. The Mysterious Flame: Conscious Minds In A Material World: Conscious Minds in a Material World. New York: Basic Books, 2000.
Moore, Gordon E. “Cramming more components onto integrated circuits.” Electronics, 38.8 (1965):114-117.
Natsoulas, Thomas. “On perceptual aboutness.” Behaviorism 5.1 (1977): 75-97.
Pearson, T. “IBM Watson. How To Build Your Own Watson Jr In Your Basement.” IBM Developer Works Blog. 2011. February. Available at: https:// www.ibm.com/developerworks/mydeveloperworks/blogs Accessed: 15 January 2016.
Russell, Stuart J. and Norvig, Peter. Artificial Intelligence. A Modern Approach. USA: Prentice Hall. 2010.
Saifi, S; Taylor, A.J; Allen, H; Hendel, R. “The Use of a Learning Community and Online Evaluation of Utilization for SPECT myocardial perfusion imaging.” Journal American College of Cardiology Imaging. 6 (2013): 823-829.
Samadi, David B. History of Robotic Surgery. Available at www.roboticoncology.com/history/ Accessed 22 February 2016.
Schank, Roger C and Abelson, Robert P. “Scripts, Plans, Goals and Understanding; An Inquiry into Human Knowledge Structures.” The American Journal of Psychology. 92.1 (1979): 176-178.
Searle, John. “Minds, Brains and Programs.” Behavioural and Brain Sciences. 3(1980): 417-424.
Searle, John. “Intrinsic Intentionality.” Behavioural and Brain Sciences. 3 (1980): 450-456.
Simon, Herbert. Models of Man: Social and Rational. New York: John Wiley and Sons Inc. 1957.
Swanson, D.B; Feltovich, P.J; Johnson, P.E. “Psychological Analysis of Physician Expertise: Implications for Design of Decision Support Systems.” In Shires, Wolf H. (ed.). Proceedings of the Second World Conference on Medical Informatics. New York: North Holland. (1977): 161-174.
Turing, Alan M. “Computing machinery and intelligence.” Mind 59.236 (1950): 433-460.
Vinge, Vernor. “The Coming Technological Singularity.” VISION – 21 Symposium. 1993 March.
Weis, S.M; Kulikowski, C.A; Amarel, S; Safir, A. “A model –based method for computer-aided medical decision making.” Artificial Intelligence. 11 (1978): 145 -172.