**1. Introduction**

The human history is full of examples where new inventions have created a significant disruption, dividing people into three broadly defined groups – proponents or early adopters, those who oppose, and those who are ambivalent [1]. When looking back at the relatively recent history of the great industrial revolution in Europe, it was not uncommon for opponents to attack and destroy new factories and new machines, with the perpetrators believing that the technological advances would eventually lead to the loss of their jobs and even entire professions [2]. As recently as in the mid-1980s, a group of mathematics teachers held a protest against the use of calculators in schools [3]. Fast-forwarding to today, calculators are now an integral part of our students' mathematics armamentarium!

Not surprisingly, our approach to artificial intelligence (AI) seems to be following a similar path. It is probably fair to say that most people are not fully aware of current (and thus future) benefits, limitations, and threats related to AI. Within medicine in general, there is little awareness of what AI actually entails, and what it is capable of at this time. It is this current state that serves as our "starting point" in the emerging debate on AI in medicine, including its integration, projected influence, and a variety of other considerations that are not dissimilar to past technology adoption paradigms.

In a very gradual and stealthy way, artificial intelligence (AI) and machine learning (ML) are becoming part of our everyday lives. This "randomly systematic" adoption process is putting humanity face-to-face with something never previously directly known to our civilization – An intelligence that may (and likely will) exceed our own. With the advent of modern computing capabilities, AI has evolved to a point where it can be integrated into everyday applications. Not surprisingly, it has been gradually introduced into various subdomains within the healthcare industry in recent years [4]. As a result, we will likely see significant shifts in education, clinical treatments and approaches, stakeholder expectations, and responsibilities – both in

terms of type and scope, as well as potentially redefinition of jobs and other typical employee characteristics across the healthcare space [4, 5].

In this chapter, we will focus on some of the most profound challenges facing humanity as the "human-AI relationship" approaches the so-called "technological singularity" – A term based on the astrophysical concept of "black hole" that denotes a point beyond which there is "no way back" to the previous state of affairs. In the case of a "black hole," the gravitational force surrounding the super-massive object becomes so powerful that not even light can escape. In the case of AI, "singularity" refers to a point where "AI-based technology" is sufficiently evolved to essentially take over "control operations" of the human civilization [6, 7]. Alternative views describe both "integration" and "plurality" as possible scenarios, where humans and AI either co-exist synergistically (plurality) [8, 9] or even integrate successfully (e.g., human-machine hybridization) [10, 11].

As we explore the realities of this new world, with omnipresent AI and the growing need for human adaptation, change and caution, the issues at hand will likely become less and less "technological" but will rather gravitate toward the ethical and spiritual domains.

#### **1.1 Destructive potential of AI**

There are many science fiction movies highlighting the potential dangers of improperly implemented AI – just a few examples of such messaging include the "Terminator" series, "Star Trek Voyager," "The Matrix" trilogy, and "I, Robot." The most common themes across these artistic works include machines "taking over" for humans as a form of "misguided stewardship," the objectivization of humans as "destructive and dangerous" followed by active control efforts, and finally, the most extreme form of "AI dominance" where highly evolved AI "machine hives" determine that humanity needs to be eliminated in its entirety [12–16].

At a less physically destructive level, but perhaps equally problematic in its extent and implications, the use of AI / ML in misdirected "societal control" efforts may represent another formidable challenge. For example, what would be required to stop a malignant governmental and/or regulatory entity that possesses powerful AI /ML tools, combined with omnipresent social media platforms, from abusing the tremendous power to misinform, manipulate, and eventually subdue entire populations? [17–22]. Such concerns have been highlighted by Elon Musk and associates in their open letter, "Pause Giant AI Experiments," which now has nearly 6000 signatories from all walks of life [23]. This open letter is blunt in stating that AI systems with humancompetitive intelligence pose a risk to our civilization. It goes on further to emphasize that the risks of powerful AI systems are likely unmanageable at this time. Appropriate oversight, tracking and regulatory frameworks must be put into place. The letter specifically addresses the inception of systems that are substantially more powerful that the more widely known GPT-4 (generative pretrained transformer 4) [24].

A further concern involves AI being used in the political and social influence spheres. Here there are ominous tidings regarding truth and transparency. In one instance, researchers at Stanford University examined whether AI could influence citizens regarding political issues such as assault weapons, a carbon tax, and parental leave…and it certainly did [25]. When looking at the case of ChatGPT in the setting of higher education, the AI-based system was able to readily pass exams at prestigious law and business schools, not to mention the United States Medical Licensing Examination [26–29]. Such issues are more than concerning, and they clearly constitute potential threats to civilized society and the ascent of man.

*Introductory Chapter: Artificial Intelligence in Healthcare – Where Do We Go from Here? DOI: http://dx.doi.org/10.5772/intechopen.111823*

#### **1.2 Potential benefits of AI**

As much as there are negatives to wider AI implementation, there are certainly amazing potential benefits that can be derived from properly harnessed AI capabilities, from reaching previously unimaginable levels of efficiency within our established processes and workflows, to human-AI hybridization that could actively enable much longer functional (and meaningful) longevity, new disease cures, and solutions to both physical and mental disability [11, 30]. Furthermore, AI will allow the simulation of unusual or theoretical situations such as legal cases before judges and negotiations with business competitors. In fact, there are multiple domains in which AI will benefit us (**Table 1**) [31]. There are also estimates that AI-driven innovation may contribute nearly \$13 trillion dollars to the world's economy by 2030 [32].

Perhaps the most significant benefits of AI will be realized in medicine and healthcare. Today's society faces staffing shortages of healthcare professionals [33–35]. In absence of sufficiently staffed healthcare organizations and institutions, current healthcare practitioners will require support to achieve optimal patient safety and levels of care provision. To this end, machine learning and AI-based systems will offer the potential to improve monitoring and alerting of healthcare providers such that patients who are in the greatest need are appropriately resourced.

It is important to note that although machine learning and AI are used interchangeably, they are not one and the same. Machine learning involves data science and the development of models based on large datasets to serve a particular function, for example, supporting diagnosis [36], time series prediction of therapeutic set points, predicting patient outcomes [37] such as readmission [38] or mortality [39]. AI, refers to an artificially intelligent system in that it can 'think and act' on its own with some degree of autonomy. While machine learning is closely related to AI, machine learning feeds into AI-based systems to leverage its results to perform some autonomous tasks. AI is best explained by looking at some of its initial use in video games, where antagonists (characters) in video games are programmed with AI to complete a primary task (e.g., stop the player from achieving some goal). AI in medicine and healthcare is the ultimate end goal, where patient data can be monitored continuously over time, and some aspects of treatment and care can be automated by AI. An example of this would be leveraging machine learning to support prediction of glucose in patients with type 1 diabetes [40] and leveraging predictions to automatically and dynamically adjust


#### **Table 1.**

*Potential benefits of AI in our lives and work [31].*

insulin delivery to maintain tight glycemic control. AI can be incorporated into this system to learn patterns in lifestyle, activity, and other pertinent variables to automatically adjust and adapt insulin delivery to further optimize glycemic control overtime given a patient's chaotic and unpredictable lifestyle. This is only one example of an application illustrating the power of machine learning and AI in healthcare.
