Artificial Intelligence (AI) refers to the ability of machines to perform tasks that typically require human intelligence. Although AI is not a new concept—dating back to the 1950s—it only became a practical tool within the past two decades. With the rapid development of three key components of modern AI—big data, computational power, and AI algorithms—numerous applications have been developed to improve human life. In dentistry, AI is now being applied across various fields, such as restorative dentistry, periodontics, orthodontics, oral and maxillofacial surgery, and prosthodontics. Most AI applications in dentistry focus on image-based diagnosis using radiographs or optical scans. However, due to limitations like the unavailability of consistent data and the computational power needed to manage 3D data, AI applications are less developed in areas beyond image-based tasks.

Evidence-Based Dentistry (EBD) is considered the gold standard for decision-making among dental professionals. At the same time, Machine Learning (ML), a subset of AI, learns from human expertise and can serve as a valuable tool in helping dental professionals at various clinical stages. This article explores the history and classification of AI, its applications in dentistry, and the relationship between EBD and ML, aiming to help dental professionals better understand AI as a tool to enhance their efficiency.

The Fourth Industrial Revolution and AI in Dentistry

The Fourth Industrial Revolution, also known as the digital age, has brought AI as one of its most significant breakthroughs. With the increasing development of electronic devices, AI analyzes the data collected by these devices, helping to enhance human life. AI can learn from human expertise and perform tasks that typically require human intelligence. One of its definitions is the theoretical research and development of computer systems capable of performing tasks such as visual perception, speech recognition, decision-making, and translation between languages—tasks that usually require human intelligence.

AI has been adopted across various industrial fields such as robotics, autonomous vehicles, smart cities, and financial analysis, among others. In medicine and dentistry, AI has found applications in medical and dental imaging, diagnostics, decision support, digital medicine, drug development, wearable technologies, hospital monitoring, and robotic and virtual assistants. In many cases, AI can help dentists and doctors reduce their workload.

Beyond diagnosing diseases using a single data source, AI can learn to integrate multiple data sources to arrive at diagnoses beyond human capabilities. For example, by combining fundus images with other medical data such as age, gender, BMI, smoking status, blood pressure, and diabetes risk, AI can predict heart disease. Therefore, AI can not only detect eye diseases like diabetic retinopathy but also predict heart disease from fundus images. The rapid progress in computational capacity (hardware), algorithmic research (software), and vast data input databases is what drives these successes.

History of Artificial Intelligence

AI is not a new term. Alan Turing, in his 1950 paper “Computing Machinery and Intelligence,” discussed the possibility of machines thinking. He speculated that by the end of the 20th century, the use of words and educated thought would evolve to the point where people could talk about thinking machines without disagreement. At that time, no specific term existed to define AI, but Turing described it as “machine intelligence.” He proposed the idea that machines could, like humans, use available information and make inferences to solve problems. Turing introduced a test to evaluate whether a machine had reached human-level intelligence, known today as the “Turing Test.”

In 1955, the term “Artificial Intelligence” was coined during a two-month workshop known as the “Dartmouth Summer Research Project on Artificial Intelligence.” However, at the time, the concept remained theoretical due to several limitations, including the lack of storage capabilities in pre-1949 computers, high costs of computer systems, and conservative views on funding this emerging field.

AI in Dentistry

In recent years, AI has begun to flourish in the field of dentistry, with applications in diagnosis, decision-making, treatment planning, and outcome prediction. Among all AI applications, diagnosis is the most common use in dentistry. AI can provide more accurate and efficient diagnoses, thereby reducing dentists’ workload. On one hand, dentists increasingly rely on computer programs for decision-making. On the other hand, these programs are becoming smarter, more precise, and more reliable for dental use.

While numerous studies and publications focus on AI in dentistry, comparing these studies in terms of study design, data allocation, and model performance (accuracy, sensitivity, and specificity) remains a challenge.

AI in Restorative Dentistry

Traditionally, dentists diagnose dental caries through visual and tactile examinations or radiography. However, diagnosing early-stage lesions, deep fissures, and secondary caries can be challenging, and many lesions are only detected when the decay is advanced, requiring more complex treatments such as crowns, root canals, or even implants. Despite the reliability of radiographs, caries diagnosis still relies heavily on the dentist’s experience.

Research in restorative dentistry focuses on diagnosing dental caries, vertical root fractures, apical lesions, volumetric pulp space assessment, and tooth wear evaluation. AI, particularly ML, learns from patterns in image features and can predict dental issues like caries. Studies have shown that AI-based detection of proximal caries outperforms traditional methods, offering more cost-effective and accurate results.

AI in Periodontics

Periodontitis is a common disease affecting billions of people worldwide. If untreated, it can lead to tooth displacement and even tooth loss. Early diagnosis and treatment are crucial to preventing periodontitis progression. Clinical diagnosis typically relies on probing depth and gum recession analysis. While the Periodontal Screening Index (PSI) is often used to assess clinical attachment loss, it is still heavily dependent on the dentist’s expertise, and parts of the periodontal tissue may be overlooked.

AI has proven to be a reliable tool for diagnosing and classifying periodontal diseases.

AI in Orthodontics

Orthodontic treatment planning is often based on the orthodontist’s experience and preferences. Since each patient and orthodontist is unique, treatment is mutually decided. Traditionally, orthodontists have had to evaluate numerous variables in cephalometric analysis to diagnose malocclusion, making treatment planning and outcome prediction complex. AI offers an ideal solution to orthodontic problems, from simulating facial changes before and after treatment to assessing skeletal patterns and anatomical landmarks on lateral cephalograms. AI can significantly aid in effective communication between patients and dentists.

Conclusion

AI has demonstrated that it can learn beyond human expertise, making it a valuable tool in dentistry. The successful development of AI relies heavily on advancements in computational technology (hardware), algorithms (software), and vast databases of input data. Tasks that involve 3D models require high computational power to train algorithms. While current computational capabilities may not be sufficient for some 3D classification tasks, ongoing developments in wearable technologies help in gathering large-scale medical data, contributing to the evolution of AI applications in dentistry.

Both ML and EBD have their advantages and limitations. While ML offers a faster approach to utilizing existing data for desired outcomes, EBD traditionally relies on controlled clinical trials to achieve its goals. Nonetheless, the integration of ML with EBD could further enhance the discovery of hidden relationships between medical data and diseases, leading to better, more personalized diagnoses for patients.