Artificial Intelligence in Medicine and Dentistry

Abstract

Introduction

Artificial intelligence has been applied in various fields throughout history, but its integration into daily life is more recent. The first applications of AI were primarily in academia and government research institutions, but as technology has advanced, AI has also been applied in industry, commerce, medicine and dentistry.

Objective

Considering that the possibilities of applying artificial intelligence are developing rapidly and that this field is one of the areas with the greatest increase in the number of newly published articles, the aim of this paper was to provide an overview of the literature and to give an insight into the possibilities of applying artificial intelligence in medicine and dentistry. In addition, the aim was to discuss its advantages and disadvantages.

Conclusion

The possibilities of applying artificial intelligence to medicine and dentistry are just being discovered. Artificial intelligence will greatly contribute to developments in medicine and dentistry, as it is a tool that enables development and progress, especially in terms of personalized healthcare that will lead to much better treatment outcomes.

Introduction

There is probably no human being in the world that has not, at some point in his or her life, become aware of the limits of his or her physical and/or mental abilities. In the past, these limits often meant life or death because there was no way to go beyond one's limits and simply improve one's abilities. The devices and machines invented and manufactured by man have undoubtedly made everyday life easier, especially in the physical sense. However, to combine the endurance and reliability of machines with intelligence and consciousness as human characteristics was the desire of many inventors, researchers and philosophers who tried to describe the process of human thinking as a mechanical manipulation of symbols. A thinking artificial creation, i.e., a machine that has a kind of meta-consciousness and thinks like a human, is something that captures the imagination. The theoretical foundations of what we now call artificial intelligence (AI) were laid by Alan Turing, Claude E. Shannon, and Norbert Wiener (1). Alan M. Turing (1912 – 1954) is an English mathematician who is considered the father of theoretical computer science; Claude E. Shannon (1916 – 2001) is an American mathematician who is known as the father of the so-called "information theory"; and Norbert Wiener (1894 - 1964) is an American mathematician and philosopher, the founder of cybernetics. They are responsible for the concept of creating intelligent machines (2). However, the concept of AI itself is somewhat younger. The concept of AI dates back to 1956, when a group of researchers participating in an eight-week Dartmouth Summer Research Project on Artificial Intelligence at Dartmouth College in New Hampshire, USA, proposed a research project and set the goal of creating "thinking machines" that could mimic human intelligence and behavior. This is widely regarded as the beginning of AI as a formal field of study (3, 4).

In order to better understand the concept of artificial intelligence, it is necessary to clarify the difference between artificial intelligence, deep learning, machine learning and data science, Figure 1. Artificial intelligence, deep learning, machine learning, and data science are related but distinct fields. Artificial intelligence is the broadest field, of which machine learning and deep learning are subsets. Data science uses techniques from all of these fields to gain insights and knowledge from data. Artificial intelligence is a broad field that encompasses a range of techniques and methods aimed at creating intelligent machines that can perform tasks that normally require human intelligence, such as visual perception, speech recognition, decision making, and natural language processing. Artificial intelligence can be divided into several branches, including expert systems, robotics, and natural language processing, to name a few. Deep learning is a subfield of artificial intelligence that uses neural networks inspired by the structure of the human brain to learn from large amounts of data. Deep learning algorithms can automatically identify and extract features from raw data such as images, sounds, and text and use them to make predictions or decisions. Examples of deep learning applications include image recognition, speech recognition, and natural language processing. Machine learning is a subfield of artificial intelligence that focuses on developing algorithms and statistical models that allow computers to learn from data without being explicitly programmed. Machine learning techniques can be supervised (the algorithm learns from labeled data), unsupervised (the algorithm learns from unlabeled data), or semi-supervised (the algorithm learns from a combination of labeled and unlabeled data). Applications of machine learning include recommendation systems, fraud detection, and predictive modeling. Data science is an interdisciplinary field that combines statistical and computational techniques with domain-specific knowledge to gain insights and knowledge from data. Data science encompasses a range of activities including data acquisition, cleaning and pre-processing, exploratory data analysis, statistical modeling, and machine learning. Data science is used in a variety of fields, including healthcare, finance, social media, and e-commerce.

Figure 1.

Artificial intelligence, deep learning, machine learning and data science

Artificial intelligence has been applied in various fields throughout history, but its integration into daily life is more recent. The first applications of AI were primarily in academia and government research institutions, but as technology has advanced, AI has also been applied in industry and commerce (5). One of the earliest examples of AI in everyday life was the use of expert systems in the 1980s and 1990s. These were computer programs that could mimic the decision-making abilities of a human expert in a particular field, such as medicine or finance. In the 21st century, AI has been increasingly integrated into a variety of consumer products and services. Examples of AI include virtual personal assistants such as Apple's Siri and Amazon's Alexa, and recommendation systems used by companies such as Netflix and Amazon to personalize their customers' experiences. AI is also being used in areas such as self-driving cars, healthcare and finance.

Today, we can distinguish three generations of AI (3). The first generation is the so-called artificial narrow intelligence (ANI), the second is the artificial general intelligence (AGI) and the third, currently the most advanced generation, is the artificial superintelligence (ASI). From today's perspective, we can safely say that the first generation is already ubiquitous. When we talk about the first generation of AI, we mean, among other things, the facial recognition and tagging technology used by Facebook, for example, virtual voice assistants in our cell phones such as Siri, Alexa or Bixby, the technology developed by Tesla and Google for self-driving cars, and much more. The second generation of AI is expected to think, plan, and solve problems and tasks on its own. The third generation of artificial superintelligence will be a truly self-aware system that could make humans and their input obsolete in certain situations. Although it is a fictional AI-based system on AI, HAL 9000 from Arthur C. Clarke's A Space Odyssey is perhaps the most vivid representation of the capabilities of third-generation artificial intelligence at this time.

Considering that the possibilities of applying artificial intelligence are developing rapidly and that, according to PubMed data, this field is one of the areas with the greatest growth in the number of newly published articles (6), the aim of this paper was to provide an overview of the literature and to give an insight into the possibilities of applying artificial intelligence in medicine and dentistry. The aim was to discuss in details its advantages and disadvantages. For this purpose, the databases Pubmed and Web of science were searched. The keywords artificial intelligence, medicine, healthcare, dentistry, deep learning, and machine learning were used.

Advantages and disadvantages of artificial intelligence in daily life

AI is expected to become even more prevalent in everyday life in the future as technology continues to advance and more industries adopt AI-powered systems and products. The use of AI in everyday life has several advantages, including (2, 7, 8):

• Improved efficiency and productivity: AI can automate repetitive tasks, making them faster, more accurate, and less error-prone.

• Personalization and customization: AI can learn from data and preferences and provide personalised recommendations, services and experiences.

• Increased security and safety: AI can be used to monitor and protect against potential threats and hazards, such as fraud, cyberattacks, and natural disasters.

• Better decision making: AI can analyse large amounts of data, identify patterns and insights, and help with decision making in various fields such as finance, healthcare, or transportation.

• Improved human capabilities: AI can complement human capabilities by providing real-time information, suggestions, and assistance.

• Improved accessibility: AI can provide services, information, and entertainment for people with disabilities or limited mobility.

• Cost efficiency: AI can help reduce the cost of many tasks and services by automating them, reducing the need for human labour.

• Predictive capabilities: AI can analyse and learn from historical data and make predictions about future events, which can be used in various fields such as weather forecasting, finance, and healthcare.

Although the application of AI has numerous advantages, facilitations and optimizations in everyday life, it also has some disadvantages (5, 9). The disadvantages of using AI in everyday are as follows:

• Job displacement: AI can automate many tasks, making them faster, more accurate, and less error-prone, which can lead to job displacement and unemployment.

• Bias and discrimination: AI systems can perpetuate and even reinforce biases and discrimination present in the data on which they are trained.

• Privacy and security concerns: AI systems require large amounts of data to function, which can lead to privacy and security concerns.

• Lack of transparency and accountability: it can be difficult to understand how AI systems make decisions, which can make it difficult to explain or hold them accountable.

• Dependence on technology: AI can become a crutch that people rely on too much, and some people may no longer be able to complete tasks without AI assistance.

• Ethical concerns: AI may raise ethical concerns, such as autonomous weapons and decision making, job displacement, and privacy.

• Lack of human contact: AI systems lack the human touch and emotion that can be important for certain tasks, such as healthcare, customer service, and education.

• Limited understanding: AI systems may have limited understanding of context and have difficulty understanding nuances and subtleties of human language and behaviour.

AI has the potential to be dangerous in certain scenarios, particularly when it comes to the development and use of autonomous weapons and decision-making systems that lack proper oversight and regulation. In February 2017, the European Parliament adopted a report with recommendations for the European Commission on civil regulation of robotics. In response to public concerns and debate, Peter John Bentley, a computer scientist at University University London, Miles Brundage of the Future of Humanity Institute at Oxford University, Olle Haeggstroem, a professor of mathematical statistics at Chalmers University and author of the book Here be dragons, and philosopher Thomas Metzinger of the University of Mainz, wrote a comprehensive document under the auspices of the European Commission entitled Should we fear artificial intelligence? in which they discuss in detail the potential dangers posed by the use of AI (10). The potential dangers and risks that may be associated with AI can be divided into the following groups:

• Decision-making systems: AI systems used to make decisions where the stakes are high, such as in healthcare, criminal justice, and finance, can perpetuate or reinforce societal biases or discrimination if they are not properly designed, tested, and regulated.

• Cybersecurity risks: AI systems can be vulnerable to hacking and cyber-attacks, which can have serious consequences when deployed in critical infrastructure such as power grids and transportation systems.

• Autonomous weapons: AI-controlled weapons that can select and attack targets without human intervention raise concerns about accountability and ethical decision-making in the event of unintended harm or collateral damage.

• Unintended consequences: AI systems can have unintended consequences, such as job displacement and economic disruption if not properly managed and regulated.

Artificial intelligence in medicine and healthcare

The decades-long use of computers in medicine and healthcare around the world has generated huge databases with an enormous amount of different data on patients, diagnoses, medical records and linked laboratory results, radiological images, clinical images, therapeutic procedures, treatment results, and much more. The existence of such large databases has created one of the main prerequisites for machine learning and the development of artificial intelligence in all areas of medicine. The number of areas in modern medicine where AI is finding practical application is growing inexorably. It is almost impossible to present all the areas in which AI is applied in medicine today, therefore this paper will present only some of them. Only 15 or 20 years ago, the application of AI in medicine was mostly experimental and geographically limited to developed and rich countries (11–18). Today, AI can be used in a variety of ways to improve healthcare and medicine:

• Medical imaging: AI can help radiologists and other medical professionals analyse medical images, such as CT scans and X-rays, to detect and diagnose diseases (19–22).

• Diagnosis and treatment: AI can analyse large amounts of patient data to help physicians make more accurate diagnoses and personalised treatment plans. Artificial intelligence is now used in almost all areas of medicine for screening, diagnosing and treating patients, including gastroenterology and digestive disorders (23, 24), cancer screening, diagnosis and treatment (25–27), COVID-19 (28), heart diseases and failures (29), oncology (30), intensive care (31), dermatology (32) and many more.

• Drug discovery and development: AI can be used to analyse big data from genetic, chemical, and medical research to identify new drug candidates and speed up drug development (33–36). Artificial intelligence has also been used in the research of new drugs and vaccines against Covid-19 (37). The role of AI in oncology is particularly important for cancer research and for the discovery of new drugs, because by applying the principles of personalized medicine it is possible to find much better and more effective drugs more quickly (38).

• Clinical decision support: AI can help physicians and other healthcare professionals make better decisions by providing real-time information and alerts based on patient data (39–42). The progress of AI application is particularly pronounced in intensive care (31), surgery (43), oncology (44) clinical decision support in infectious diseases (45) but also in other areas of medicine.

• Personalized medicine: AI can be used to analyze genetic and patient data to create personalized medicine (46) plans and treatments tailored to the specific needs of individual patients especially in oncology (30, 47), cardiovascular medicine (48–50).

• Monitoring and tracking of chronic diseases: AI can help monitor vital signs, symptoms and other data to help identify potential health issues early (51, 52), especially in cases of chronic diseases, helping to prevent complications like in chronic obstructive pulmonary disease (53) or hypertension management (54).

• Predictive analytics: AI can help predict the likelihood of certain medical conditions and diseases based on a patient's data, thus contributing to their prevention and treatment (55, 56). Large existing public health system databases are used for this purpose (57). It has been interesting to see how AI has been used to predict COVID -19 outcomes (58) or, in patients with gastrointestinal cancer, to predict response to treatment (59).

• Medical research: AI can be used to analyse large volumes of medical data, identify patterns, and gain new insights that can help understand the pathology of diseases and ultimately develop new treatments (60–63).

Advantages and disadvantages of artificial intelligence in medicine and healthcare

In general, the advantages and disadvantages of using AI in medicine are similar to the advantages and disadvantages of using AI in daily life. The use of AI in medicine has several advantages, including:

• Improved accuracy and efficiency: AI can help doctors and other healthcare professionals analyse large amounts of medical data, such as imaging and patient records, to detect and diagnose diseases more quickly and accurately (64).

• Personalized medicine: AI can be used to analyze genetic and patient data to create personalized medicine plans and treatments tailored to the specific needs of individual patients.

• Clinical decision support: AI can help physicians and other healthcare professionals make decisions by providing real-time information and alerts based on patient data (65).

• Early detection and prevention: AI can help monitor vital signs, symptoms, and other data to detect potential health problems early, especially chronic diseases, to avoid complications (66).

• Predictive analytics: AI can help predict the likelihood of certain medical conditions and diseases based on a patient's data, helping to prevent and treat them.

• Medical research: AI can be used to analyse large volumes of medical data, identify patterns, and make new discoveries that can help understand the pathology of diseases and ultimately develop new treatments.

• Reduced costs: Using AI to automate certain tasks, such as analysing imaging and patient data, can reduce the need for manual labour, resulting in cost savings (67, 68).

• Remote care: AI can help monitor patients remotely, which can be particularly useful for people living in remote areas or those with mobility issues to enable access to healthcare services (53, 69, 70).

The use of AI in medicine has several drawbacks, including:

• Bias and discrimination: AI systems can perpetuate and even reinforce biases and discrimination present in the data on which they have been trained. This can lead to incorrect diagnosis or treatment for certain groups of people (71, 72).

• Lack of transparency and accountability: it can be difficult to understand how AI systems make decisions, which can make it difficult to explain or hold them accountable (73–75).

• Privacy and security concerns: AI systems require large amounts of patient data to function, which can lead to privacy and security concerns, especially with the growth of electronic medical records (76, 77).

• Dependence on technology: AI may become a crutch that physicians and other healthcare professionals rely on too heavily, and some may no longer be able to complete tasks without AI assistance (78).

• Limited understanding: AI systems may have limited understanding of context and have difficulty understanding nuances and subtleties of human health and disease (79, 80).

• Ethical concerns: AI may raise ethical concerns, such as autonomy and decision-making in healthcare, e.g., the use of autonomous surgical robots (81, 82).

• Job displacement: AI can automate certain tasks, such as analysing imaging and patient data, which can lead to job displacement and unemployment (83–85).

• Lack of human touch: AI systems lack the human touch and emotions that can be important for certain tasks such as healthcare, customer service, and education (86–88).

It is important that these potential drawbacks be considered in the development and implementation of AI systems in the medical field and that regulation and controls are in place to mitigate potential adverse effects. It is also important to ensure that AI systems are used as tools to assist physicians and other health care professionals, rather than replacing them.

Artificial intelligence in dental medicine

The speed with which analog dentistry is being replaced by digital dentistry is comparable to the speed with which artificial intelligence is penetrating profusely through the daily work of a modern dentist. Although these changes are not always immediately recognized or associated with AI, they are neither small nor insignificant, especially when viewed in a broader context. AI can be used in a variety of ways to improve dental care and dentistry, such as segmenting and identifying teeth (89, 90), planning of dental implants treatment, identification and classification of dental implant systems (91), for detection and classification of dental plague (92), for diagnosing maxillary sinusitis on panoramic radiography (93), for cephalometric landmarks detection (94), or for root morphological classification (95), dental caries detection on periapical and bitewing X-ray images (96), and many other applications including (97–105):

• Dental imaging: AI can assist dentists and other dental professionals in analysing dental images, such as X-rays and CT scans, to help identify and diagnose conditions such as cavities and periodontal disease (6, 103, 106–108).

• Treatment planning: AI can be used to analyse dental images and patient data to help dentists create personalised treatment plans in almost all areas of modern dentistry (101, 103, 109–112).

• Orthodontics: AI can be used to analyse dental images to create 3D models of teeth and jaws that can be used for orthodontic treatment planning and simulations (103, 113, 114).

• Dental prosthetics: AI can be used to create 3D models of teeth and jaws that can be used to fabricate crowns and bridges, but AI can also be used in dental CAD /CAM systems to help dentists design and fabricate dental restorations such as fillings, crowns, and bridges using computer-aided design and manufacturing systems (103, 115, 116).

• Periodontology: AI can be used to differentiate between aggressive and chronic periodontitis and to diagnose aggressive or chronic periodontitis using relatively simple and easy to determine parameters such as the leukocyte count in the peripheral blood (103, 117).

• Endodontics: AI has been introduced to determine the root canal morphology, locating minor apical foramen, detecting periapical lesions and root canal fractures, evaluation of success of treatment and retreatment (95, 103, 104).

• Oral pathology: AI can be a promising aid in the diagnosis of head and neck cancer lesions and offers great potential for detecting tumour tissue in tissue samples or on radiographs (103, 118–120).

• Forensic dentistry: AI can be used for dental profiling including estimation of age and determination of sex of an individual using X-ray images (121–129).

• Dental robotics: AI can help dentists perform certain procedures in dental implantology, oral and maxillofacial surgery, prosthetic and restorative dentistry, endodontics, orthodontics, oral radiology, and dental education through the use of robotic systems (130–134).

• Chatbots: AI-driven chatbots can help patients make appointments, answer questions, and educate them about dental care (135, 136).

Advantages and disadvantages of artificial intelligence in dentistry

Technological advances in medicine and dentistry bring numerous advantages that have a positive impact on maintaining or achieving oral health, but also certain problems and doubts. The use of AI in dentistry has several advantages, including (137, 138):

• Improved accuracy and efficiency: AI can help dentists and other dental professionals analyze large volumes of dental data, such as imaging and patient records, to detect and diagnose conditions more quickly and accurately (139, 140).

• Personalized treatment plans: AI can be used to analyze dental images and patient data to help dentists create personalized treatment plans tailored to the specific needs of individual patients (141).

• Predictive analytics: AI can help predict the likelihood of certain dental problems and diseases based on a patient's data, helping to prevent and treat them (142–146).

• Lower costs: using AI to automate certain tasks, such as analysis of dental imaging and patient data, can reduce the need for manual labour, leading to cost savings (147).

As in medicine, shortcomings in the use of AI in dentistry may be associated with bias and discrimination, lack of transparency and accountability, privacy and security concerns, particularly with the advent of electronic dental records, and reliance on the technology among dentists and other dental professionals. AI systems may be limited in their understanding of the context of human oral health and disease, which can cause serious problems as well as harm. Currently, job displacement cannot be considered an important shortcoming of AI application in dentistry, but it is possible that some tasks and procedures could be replaced by AI devices. Although patients are positive about AI in dentistry (148), AI in dentistry, as in medicine, may raise some ethical concerns (149), mainly related to prudence, justice, privacy, responsibility, solidarity, autonomy, and health care decision making (150). In addition, AI systems may not be accessible or affordable to all people and communities, which may lead to inequities in access to health care. It is important to note that AI in dentistry is still in progress and its utility depends on the particular use case and implementation. It is also important that adequate regulations ensure that AI systems are safe, effective, and ethical.

It is of utmost importance to consider these potential drawbacks when developing and deploying AI systems in the dental field and to have regulations which can mitigate potential negative impacts. Ensuring that AI systems are used as tools to assist dentists and other dental professionals is more important than replacing them.

Conclusions

The possibilities of applying artificial intelligence in medicine and dentistry are just being discovered. It is expected that there will be a revolution in healthcare in the coming years as there will be more and more efforts to provide personalized healthcare that will lead to much better outcomes. The main applications of artificial intelligence in medicine are: medical imaging, diagnosis and treatment, drug discovery and development, clinical decision support, chronic disease monitoring and tracking, predictive analytics, and medical research. In dentistry, artificial intelligence can be used for dental imaging, diagnosis and treatment planning in orthodontics, prosthodontics, periodontics, endodontics, oral pathology, and also in forensic dentistry for dental profiling. Artificial intelligence will be a big part of this evolution as a tool that enables development and progress. This will ultimately lead to better health; will improve one's quality of living, thus enabling longer life expectancy. Nevertheless, the shortcomings of artificial intelligence should be taken into account, especially those of ethical nature. States and individuals very easily accept the benefits that artificial intelligence brings, but they are very slow to develop and implement the rules that regulate it. This must be changed before it is too late.