The emergence and integration of artificial intelligence (AI) into various facets of clinical chemistry over the past decade have opened exciting opportunities and transformative possibilities in how we practice laboratory medicine. Clinical chemistry, as a specialized field within laboratory medicine, is no exception in this regard. AI and its associated tools have undergone extensive exploration recently to identify novel applications in the everyday practice of clinical chemistry. This pursuit has become even more crucial as the global healthcare landscape continues to adapt to the profound changes brought about by the COVID-19 pandemic, which significantly impacted all aspects of laboratory medicine [1, 2].

Artificial Intelligence (AI) is referred to as the capacity of machines to replicate human intelligence. It is a broad terminology that comprises multiple subdomains, including Artificial General Intelligence (AGI, where AI can think like humans), Artificial Super Intelligence (ASI, where AI surpasses human capabilities, a far-fetched reality at present), and Artificial Narrow Intelligence (ANI, encompassing methods from traditional machine learning to advanced deep learning, primarily applied for specific automated tasks). AGI, with its all-encompassing and task-flexible nature, is often referred to as 'strong AI,' while ANI, with its focused and task-specific applications, is termed 'weak AI.' Currently, despite being a subject of active research, AGI remains elusive, and most of today's AI applications are rooted in ANI. However, it's worth noting that ChatGPT, initially developed based on ANI, has undergone a series of modifications, and is gradually approaching AGI capabilities [3, 4].

Machine learning (ML), an ideal example of ANI, possesses the ability to process vast datasets and uncover intricate patterns that often surpass the capabilities of rule-based systems and human experts. ML can be broadly categorized into two subdomains: supervised machine learning and unsupervised machine learning. Supervised machine learning, the predominant subfield of ML in clinical chemistry, involves the computer's capacity to deduce patterns from pre-existing labelled data, where the desired outcome is already known. These labels provide feedback to the computer program, enabling it to determine the correct answers and enhancing its predictive capabilities (for example, prediction of acute kidney injury based on laboratory test results with use of previous results from patients who developed this disease). Numerous supervised machine learning techniques have been applied in clinical chemistry, including regression (linear and logistic), tree-based models (e.g., random forest, XGBoost), and support vector machines. On the other hand, unsupervised machine learning involves models working with unlabelled datasets to uncover underlying patterns or trends. This approach can reveal insights that were previously unrecognized. Common examples of unsupervised machine learning models include k-means clustering, k-nearest neighbours, and principal component analysis [3].

In the context of clinical chemistry, AI and its tools have found applications across various aspects of laboratory operations and performance. A significant area of focus involves the prediction of laboratory test values and optimizing laboratory resource utilization. Researchers have developed algorithms and models to power clinical decision support tools, which, in turn, can aid in optimizing laboratory test utilization. For instance, Azarkhish et al. developed a neural network model capable of predicting serum iron levels and identifying iron deficiency anaemia using information derived from complete blood counts [5]. Similarly, Luo et al. have reported a clinical decision support tool with the potential to predict test results from clinical data and other pertinent laboratory parameters [6]. In another study, Lee et al. introduced a deep learning-based neural network model for predicting low-density lipoprotein cholesterol (LDL-C) using lipid profile parameters, such as total cholesterol, triglycerides, and high-density lipoprotein cholesterol. Notably, they demonstrated that this model outperformed traditional methods for estimating LDL-C, such as the Friedewald equation and Martin's method [7]. AI tools have also been applied to interpret test results and establish personalized reference ranges—a development with profound implications for disease diagnosis and interpretation in clinical chemistry. Additionally, ML techniques have proven valuable in test result validation, quality control, and laboratory information systems, thereby enabling automation in test result verification, improving laboratory operations, and supporting clinical research endeavours [8].

The application of artificial intelligence in clinical chemistry holds immense promise, given the pivotal role of laboratory professionals in numerous medical decisions [9]. The field's intricate and expanding data landscape necessitates the integration of AI to effectively manage and analyse information. Clinical chemists play a critical role in safeguarding this data, underscoring the need for them to develop a comprehensive understanding of AI capabilities and recognized constraints within the laboratory medicine domain. Clinical chemistry experts, with their profound knowledge of optimal test development methodologies, should actively engage in advancing AI technologies, as there are significant parallels between AI and laboratory-developed tests. Although still in the early stages of development, artificial intelligence and machine learning are already employed to automate numerous laboratory functions, optimize resource allocation, provide personalized reference ranges, and test interpretations. Given the evolving nature of clinical chemistry, laboratory professionals must familiarize themselves with these technological advancements and prepare for their imminent integration into future practice. The existing body of research increasingly points toward the substantial utilization of these methodologies in shaping the future of clinical chemistry, enhancing its performance attributes. Acquiring foundational knowledge in artificial intelligence is bound to augment cross-disciplinary literacy, ultimately leading to improved integration and comprehension of these technologies within the field of clinical chemistry.