Alan Turing first proposed the idea of AI and conceived the term in 1950. He proposed the “imitation game” to understand the machine's capacity, which can be distinguished from humans (Turing, 1950; Zador et al., 2023). In 1966 at MIT, The first Chatbot was developed with the name “ELIZA” by Weizenbaum (1966). It was the first program that was competent in attempting the Turing test. The perspective was to act as a psychotherapist (Weizenbaum, 1966). The program applied pattern matching (Brandtzaeg and Følstad, 2017). After a few years, ELIZA was improved, and a new Chatbot was developed in 1972, which was named PARRY (Colby et al., 1971). Again another new Chatbot was created in 1995, named ALICE. It used a pattern-matching algorithm and was the first Chatbot using the AIML (Artificial Intelligence Markup Language) (Marietto et al., 2013).

Several Chatbots were developed from time to time: Apple Siri, IBM Watson, Amazon Alexa, Microsoft Cortana, and Google Assistant. The recently developed famous one was ChatGPT (Figure 3).

The essential components or technology used for Chatbots are Pattern Matching, Natural Language Understanding (NLU), Artificial Intelligence Markup Language (AIML), Latent Semantic Analysis (LSA), Natural Language Processing (NLP), Chatscript, entity, RiveScript, contexts, etc.

In pattern matching, the input text was compared with the text components of a database. It used input (stimuli) sentence and output (response). Eliza and ALICE were developed using pattern matching (Dale, 2016). AIML is an XML-based language based on pattern matching or pattern recognition concepts. It uses tag-based basic dialogue called categories (<category>). ALICE has used the AIML. There are various training methods for Chatbots, each with its own strengths and weaknesses. One common approach is the rule-based method, which uses a set of rules to guide the Chatbot's responses. This can be effective for simple use cases but may not handle complex user queries, and scaling can be difficult. Another procedure is the machine learning method, which involves training the Chatbot on a large dataset of conversations using algorithms. This method can handle complex queries and improve over time, but it requires a significant amount of training data and can be computationally expensive. The third technique is the hybrid approach which combines rule-based and ML methods for a more robust Chatbot. The Chatbot follows the rules for simple questions and uses ML for complex ones. This process can handle a wide range of user queries and improve over time; however, it requires expertise in both methods. Real-life illustrations include Google Assistant, which works on a hybrid methodology and provides users with personalized responses. It has the capability to process a vast array of user queries and can enhance its performance over time by assimilating more data. Secondly, we have Amazon Alexa, which employs ML to comprehend user queries and furnish personalized responses. It is proficient in handling intricate queries, and its performance can be optimized by learning from more data over time. Chatbots based on the three training methods mentioned above offer 24/7 availability, cost-effectiveness, and consistency in responding to user queries. However, they may lack empathy, have a limited understanding of complex queries, and are limited by technical constraints.

The essential constituents of a Chatbot system encompass a User Interface (UI), NLP, ML and DL, Dialog Management, and Integration. The UI interacts with the user and it can either be a web interface, a mobile app, or a messaging platform like Facebook Messenger or WhatsApp. NLP component, on the other hand, is responsible for comprehending the user's input and extracting relevant information. It applies techniques such as tokenization, part-of-speech tagging, and named entity recognition to evaluate the user's text. Modern-day Chatbots rely on the use of NLU and Natural Language Generation (NLG) to recognize and respond to users. This is achieved by leveraging the ML and DL elements of AI, which provide responses based on user interactions. The knowledge base of the Chatbot serves as a repository of information which it can utilize in answering user queries. This base encompasses a variety of forms, including databases, sets of rules, or corpora of text. The Dialog Management component assumes responsibility for overseeing the conversation flow and generating appropriate responses based on the user's input and context. It employs a range of techniques, namely state machines, decision trees, or DL models. Finally, the Integration component enables the Chatbot to establish connections with external systems, such as APIs, databases, or other Chatbots. The Chatbot can be developed using cutting-edge programming languages, including but not limited to C++, J, Python, Java, Lisp, PHP, Ruby, Clojure, or any other programming language with similar capabilities (Figures 4A–C).

Chatbot technology can be used in the medical industry to improve patient care by giving patients 24 h medical advice and assistance. Chatbots that simulate human-like speech are designed to help people manage their health and communicate with their healthcare providers. The workflow, monitoring, diagnosis, treatment, and promotion of good health are other areas where Chatbots might be helpful. Chatbots have made healthcare services more readily available, practical, and effective (Chow and Sanders, 2023). However, other challenges include the need for human aspects in healthcare and problems with accuracy and dependability (Xu et al., 2021). Chatbots are revolutionizing healthcare by giving patients 24/7 access to medical information, setting appointments, and gathering patient data. Healthcare Chatbots that offer exceptional, individualized care have improved patient care (Safi et al., 2020). The capacity of Chatbots to offer patients immediate support and assistance is one of the reasons why the healthcare sector is seeing an uptick in their use. The NLP, decision trees, and ML techniques can all be used to build and develop medical Chatbots (Kidwai and Nadesh, 2020). From a medical point of view, there are three broad varieties of medical Chatbots: informational, diagnostic, and administrative (Safi et al., 2020). Medical Chatbots are becoming a more popular source of health literacy among users. It can take time to assess the worth and efficiency of medical Chatbots, though, several techniques, such as user testing, surveys, and data analysis, can be used to evaluate the usefulness and usability of medical Chatbots (Mokmin and Ibrahim, 2021).

At the same time, researchers believe the healthcare sector's AI-based mode can lower the growing carbon footprint (Wolf et al., 2022; An et al., 2023; Das and Chandra, 2023). Chatbots can diminish the carbon footprint of healthcare facilities by decreasing the necessity for patients to commute to hospitals for minor concerns and uncertainties. Carbon emissions can be substantially reduced through remote consultations for each appointment. So, chatbots can result in a decline in energy-intensive processes and transportation emissions related to healthcare provision. Furthermore, predictive analytics driven by AI can optimize the allocation of resources, thus minimizing excessive utilization of facilities and resources. Moreover, AI can facilitate the early identification and treatment of diseases, potentially lessening the environmental impact of advanced-stage treatments. Consequently, healthcare models driven by AI not only enhance patient outcomes but also contribute to a more environmentally friendly and sustainable healthcare system.

Several Chatbots are applied in medical science to provide some solutions in the medical field from time to time, such as Smart Chatbot, ChatGPT, EQRbotChatbot, Ask Rosa, TanaChatbot, Vickybot, COVID-Bot, Otis Chatbot, SnehAIChatbot, Anti-TB Chatbot, and ChatbotWakamola (Table 2). GYANT, a healthcare Chatbot that diagnoses patients and administers medication, and over-the-counter (OTC) Chatbots, which recommend nutritious meals and activities, are two prominent examples of healthcare Chatbots. One of the issues with employing medical Chatbots is ethical concerns. Numerous academicians and medical professionals have emphasized that Chatbots are not technically capable of diagnosing medical issues or substituting for professional opinions (Parviainen and Rantala, 2022; Haug and Drazen, 2023). Another issue is that patients generally do not trust medical Chatbots due to low confidence in offering reliable guidance or information. Medical Chatbots may use biased algorithms that can be harmful (Sharma et al., 2022). While Chatbot algorithms can perform simple diagnostic procedures, there are some challenges and limitations. Online consultations, other forms of virtual assistance, and other crucial elements could overlook for a reliable result. Despite these challenges and limitations, medical Chatbots have an opportunity to transform many aspects of healthcare. Healthcare providers utilize medical Chatbots to respond the inquiries, create medical records and provide disease information (Chang et al., 2022).

The enormous language model-based question-and-answer system known as ChatGPT has created business and academic interest (Guo et al., 2023). The OpenAI Company in the USA developed ChatGPT. Presently, OpenAI has released ChatGPT plus which is a subscription plan for ChatGPT. OpenAI is an AI-based research laboratory that developed another famous Chatbot, GPT-4 (Lee et al., 2023; Sanderson, 2023). ChatGPT is a platform that can perform language translation and answer simple questions. On the other hand, GPT-4 is more intelligent and can comprehend images along with the text. As a state-of-the-art AI system, ChatGPT can produce content with a natural conversational flow and react to requests fast. As a result, for example, poetry, children's books, and fan fiction can all be created. ChatGPT has undergone considerable testing, and the results have been positive in various industries, including software testing, education, and healthcare (Roberts et al., 2020). In-depth human evaluations and linguistic analyses of ChatGPT-generated content in comparison to that produced by people have been carried out by researchers to understand the traits of ChatGPT's replies, better the differences and gaps from human experts, and the future directions for huge language models (Anderson et al., 2023; Jalil et al., 2023; Jang and Lukasiewicz, 2023). The capability of ChatGPT to generate text that resembles human-written language may revolutionize medical practitioners' interactions with patients, data analysis, and research. ChatGPT's integration into the healthcare system and its opportunities and challenges in medicine have been highlighted (Hügle, 2023; King, 2023; Loh, 2023).

NLP, a subfield of AI that evaluates human language, has been used in several medical specialties. In ophthalmology, NLP has extracted specific language from free-text notes for quantitative analysis, such as visual acuity (Chen and Baxter, 2022). In databases of clinical notes, pathology reports for diabetes, schizophrenia, cancer, and cardiovascular illness have been analyzed using ML-based NLP. NLP has been used extensively, whether using DL or a conventional method (Mollaei et al., 2022). In the study and therapy of breast cancer, NLP has been used to interpret pathology reports, discover penetrance records for breast and other cancer susceptibility genes, and mine electronic medical data (Hughes et al., 2020). NLP has been used in traditional Chinese Medicine to extract information from unstructured text data (Zhang et al., 2022). Clinical neurosciences have used NLP easily to synthesize literature, extract data, identify patients, automate clinical reporting, and forecast outcomes. By utilizing the enormous amount of text data now available through big data analytics and predictive AI, NLP offers the potential to advance healthcare (Zhang et al., 2022).

The architectural foundation of ChatGPT is a transformer model, a class of neural networks that excels at tasks requiring NLP. The capacity of ChatGPT to use DL techniques in producing responses to text input, that are human-like and far superior to those of any other AI model has attracted a lot of attention (Wen and Wang, 2023). It has a unique architecture in terms of processing based on the neural network that handles sequential data. It uses DL-based language models. The models have been designed to process and generate responses using NLU and NLG for any sequence of characters that make meaning from various spoken languages, mathematical equations, and programming languages (Figure 5). A significant volume of material from books, journals, and websites is used to train the model using unsupervised learning techniques. The user interface is connected with middle-layer technologies. However, the model learns to anticipate the next word in a sentence based on the words before it during training. This method allows the model to learn the significance and organization of language (Tampuu et al., 2020). The capacity of ChatGPT to respond logically and contextually suitable to various text questions is one of its primary features. Sophisticated language modeling methods and a sizable and varied training dataset are utilized to perform this function. By training ChatGPT on a smaller and more targeted dataset, specific tasks like sentiment analysis or question-answering can be made better (Iversen et al., 2014). ChatGPT's progress is significant for the study of NLP in general, as it has the potential to radically change how consumers interact with computers and other digital devices. Language models may grow even more sophisticated in the future, given how constantly evolving language model design and training methods are (Cherubini and Dinh, 2023; Shoaib et al., 2023).

Medical diagnosis and treatment are among the most significant application areas for ChatGPT. But whether it is trustworthy and truthful in these fields is still debatable. Numerous researches have examined the effectiveness of ChatGPT in treating patients and identifying illnesses (Huh, 2023). After studying ten published case reports from Korea, ChatGPT successfully identified three cases when given the patient's symptoms, findings, and medical history in a single examination (Hariri, 2023). After adding laboratory, pathological, and radiographic data, the accuracy rate rose to seven out of ten. ChatGPT was employed in a different study to provide information on cirrhosis and hepatocellular cancer (Yeo et al., 2023). Even though the questions and answers contained a lot of information regarding both illnesses, it was discovered that only a few were regarded as complete. ChatGPT still has issues, such as a tendency for biased responses and the ability to perpetuate undesirable linguistic patterns, despite having the potential to aid medical decision-making and instruction. Therefore, more research is required to assess its validity and precision in medical diagnosis and therapy (Huang et al., 2023b).

A recent experiment published in JAMA tested ChatGPT based on GPT-4 for diagnostic accuracy in complex medical cases. The findings indicate that ChatGPT can accurately identify correct diagnosis in almost 40% of cases, as its primary suggestion, and also deliver the correct diagnosis response as one of its potential advances in two-thirds of the remaining cases tested (Meck, 2023). In another study, Liu et al. (2023) demonstrated the efficacy of ChatGPT in producing clinical letters, radiology reports, medical notes, and documentation for clinical purposes. They observed that it could also furnish automated scoring, instructional assistance, enhanced individualized learning, research aid, generation of clinical scenarios, swift information retrieval, translation, clinical decision support, and personalized medicine. Further, ChatGPT provided a higher percentage of factual and accurate responses, thereby rendering it a more secure and precise instrument for various tasks. Similarly, Ali et al. (2023) have informed researchers in the digital health field that ChatGPT can prepare a clinical letter for the patient. Although ChatGPT provides vast information about medical diagnosis and treatment, several scientists have raised the question about the misinformation provided by ChatGPT (De Angelis et al., 2023). However, more accurate, next-generation, error-free ChatGPT is needed in medical diagnosis and treatment. The domain-specific next-generation, error-free ChatGPT might solve the problem in the future (Chatterjee et al., 2023; Pal et al., 2023).

Along with medical diagnosis and treatment, medical service providers are trying to use ChatGPT to aid medical laboratories, find potential research topics, and inform medical students, doctors, nurses, and other healthcare professionals about updates and new developments in their fields (Dave et al., 2023; Jeblick et al., 2023; Sedaghat, 2023). It has been used in different areas of medical science by researchers. The applied areas are clinical toxicology, medical imaging, medical education, clinical management, cirrhosis and hepatocellular carcinoma, medical report generation, diabetes research, etc. (Table 3). ChatGPT has been used in other fields, such as essay writing, programming, code debugging, teaching, etc. (Table 4). Now, ChatGPT provides medical domain-specific information. Cocci et al. (2023) informed that ChatGPT provides outcomes for urology patients. They performed a case study using 100 patients' data. They found ChatGPT provides interactive information with 52% correct answers. ChatGPT-derived answers were more pertinent for non-oncology areas (58.5%) than oncology (52.6%). Similarly, Singh et al. discussed how it impacts the ophthalmological area. They illustrated that ChatGPT can write ophthalmic discharge summaries with meaningful generic text. The discharge summary can include the medications, next consultation time, follow-up instructions, etc. (Singh et al., 2023). Simultaneously, Dossantos et al. (2023) wrote an editorial discussing ChatGPT's limitations and applications in ophthalmology. In the same direction, medical researchers discussed ChatGPT's impact on orthopedics, radiology, etc. (Fayed et al., 2023; Jiang et al., 2023).

ChatGPT or other Chatbot technologies help pharmaceutical companies provide information on drug discovery and development, and also other therapeutic areas. One recent article by Savage provides information on how ChatGPT is helping scientists to produce probable drug targets and, thereby, offering helpful information about drug targets. This AI-driven platform is a new way to target discovery by augmenting the integration and relationships provided through the knowledge graphs (Savage, 2023).

At the same time, it has been noted that ChatGPT-derived information helps to identify possible hit compounds for drug discovery. This Chatbot can help identify disease-specific genes, agents, and other potential information. It can predict the pharmacodynamics (PD), pharmacokinetic (PK), and toxicity character of a specific compound, which is extremely helpful for drug discovery and development (Zhao and Wu, 2023). Recently, Heck described that ChatGPT could provide cellular information like HSP70, which might be extremely helpful for drug discovery and development (Heck, 2023). At the same time, researchers are trying to develop the domain-specific Chatbot for drug discovery and development. One example is DrugChat, which can produce molecular graphs of drugs (Liang et al., 2023). It is envisaged that ChatGPT or other Chatbot technologies will help to revolutionize drug discovery and development shortly.

Clinical practices, medical education, and research could all be transformed by using ChatGPT. When utilized in medicine, it brings up possible ethical issues that must be resolved. ChatGPT needs extensive data to enhance its language generation abilities, but this information often carries sensitive personal details, such as medical records, that require safeguarding to ensure privacy (Ovalle et al., 2023). There are extra safety and security issues to consider because the model might not be able to handle sensitive medical information properly. More reliance on technology is needed since practitioners may stop using their clinical discretion instead of the model's suggestions (Pennestrı̀ and Banfi, 2022; Corsello and Santangelo, 2023). Ethical concerns include the need for informed consent, the likelihood that the model may replace real doctors leading to job losses, and a lack of human interaction in healthcare. However, in a practical scenario, AI can never replace real doctors, lacking human interpretation, emotion, and a sense of accountability. Another ethical concern involves potential biases that may occur in healthcare and medicine. ChatGPT's output can be compromised if the training data is biased toward a particular demographic group. Also, using ChatGPT and other AI tools in medical writing raises ethical and legal issues, namely, a potential violation of copyright laws, complications in medico-legal matters, and the requirement for transparency in AI-created content (Liu et al., 2023). Therefore, it is very clear that the replaceability of human physicians is limited, as they bear huge responsibilities. AI cannot be a substitute for a human doctor. Understanding the potential hazards, restrictions, and ramifications of these technologies' use in the medical sector is crucial for addressing these challenges. Some recommendations and good practices can mitigate the current ethical concerns. These include ensuring appropriate measures to protect patient privacy when utilizing ChatGPT and one such measure is utilizing text ambiguation frameworks to preserve user privacy. The next one is ensuring that the training data utilized to train ChatGPT is diverse and representative of the population to avoid biases in its language generation. Ensuring proper legal measures to safeguard patient privacy while using ChatGPT in healthcare and medicine is also essential. ChatGPT should follow the norms set up by governments worldwide, like HIPAA and GDPR.

Furthermore, access to these technologies could be limited in various ways, like by subscription fees, etc., to create novel forms of inequality and discrimination. When using ChatGPT professionally for academic writing, it is important to exercise caution to avoid socially sensitive content, misinformation, or plagiarism. A detailed explanation of the AI methodology should be provided to highlight the contribution of AI in medical writing or research. This includes AI's ML and DL algorithms, specific model architecture, and any alterations or adjustments made to the AI system. The method should be trained and validated as an AI model. Information on how the model's performance should be measured, such as cross-validation or separate test sets, should also be provided. Data utilized to train and test the AI system must be described, specifying the dataset's source, size, and characteristics and any preprocessing steps taken, such as data cleaning, normalization, or feature engineering. Additionally, ensuring transparency in AI-generated content and avoiding any violation of patents is crucial. At the same time, science should go on, and AI-based Chatbot technologies should be used appropriately for the sake of human society. Therefore, these technologies should be used safely, productively, and outcome-based. The publishers, researchers, and countries' regulatory authorities need to sit together and lay down rules about the use of LLMs ethically. One editorial in Nature recently illustrated the role of scientific publishers in deploying AI in medical writing (Editorial, 2023a). The development of AI language models has substantially advanced AI and has the potential to disrupt current clinical practices in all surgical and clinical specialties of medicine. These technologies' ethical and social implications must also be considered to ensure their responsible and ethical use (Aggarwal et al., 2021; Saedi et al., 2021; Lennon et al., 2022; Pennestrı̀ and Banfi, 2022; Editorial, 2023a).

In one study, NLP was used to predict whether candidates for residency interviews would receive an invitation (Mahtani et al., 2023). The NLP model's average AUROC score of 0.80 indicates its predictability. Another study used a hybrid modeling strategy to abstract CT imaging indications by combining structured data from electronic health records with NLP from radiology reports. The study found that for CT imaging indications, the NLP model could accurately distinguish between surveillance and other causes (Khan et al., 2023a). On the other hand, ChatGPT has been utilized in medical research to produce text-based answers to inquiries. Despite being able to deliver responses that resemble those of a human, it has issues with accuracy and bias. For instance, a recent study discovered that ChatGPT produced biased replies to questions about mental health.

Additionally, ChatGPT needs domain-specific knowledge, so it could not be as accurate as other NLP models in analyzing medical data (Dahmen et al., 2023). Appropriate evaluation metrics and assessment techniques must be devised to measure ChatGPT's efficacy (Vat et al., 2021). In a study, ChatGPT's performance was evaluated in the medical physiology examination of phase I MBBS (Subramani et al., 2023). The accuracy and applicability of the information provided in ChatGPT's responses were assessed using theoretical questions from the study. The findings demonstrated that ChatGPT performed well in giving accurate and pertinent information. ACGME milestones have been employed as evaluation measures in emergency medicine (Salzman et al., 2018). Another study examined three specific evaluation measures to assess clinicians' competence using high-fidelity simulation as an assessment tool (Leuck et al., 2013). Evaluation metrics are crucial for determining whether AI and ML are valuable tools. ML and AI are becoming increasingly popular in radiology and medicine, and the number of studies in this area is overgrowing (Handelman et al., 2019).

One of the case studies illustrating the potential of ChatGPT in the healthcare domain pertains to the benchmarking of ChatGPT-4 on the ACR Radiation Oncology In-Training (TXIT) Exam and Red Journal Gray Zone Cases. This particular study aimed to assess the performance of ChatGPT-4 in the specialized field of radiation oncology by employing the 38th ACR TXIT exam and the 2022 Red Journal Gray Zone cases. ChatGPT-4 attained scores of 63.65% and 74.57% on the TXIT exam, thereby underscoring the advantages offered by the latest iteration of the ChatGPT-4 model (Huang et al., 2023a). Furthermore, this study effectively identified the areas in which ChatGPT-4 excelled and displayed weaknesses within the realm of radiation oncology, thus showcasing its potential in medical education for the general public and in facilitating decision-making in the field of radiation oncology. Another noteworthy case study by Ma et al. (2023) pertains to evaluating ChatGPT's capabilities in medical report generation. This particular study proposed the employment of ImpressionGPT, an iterative optimizing framework for radiology report summarization in conjunction with ChatGPT. Remarkably, the proposed model achieved state-of-the-art performance on both the MIMIC-CXR and OpenI datasets without requiring supplementary training data.

Consequently, this study effectively demonstrated the potential of ChatGPT in the context of radiology report summarization. This development could enhance radiologists' efficiency while facilitating improved communication between radiologists and other medical practitioners. Gilson et al. (2023) investigation appraised the efficacy of ChatGPT in addressing questions encompassed within the United States Medical Licensing Examination (USMLE) Step 1 and Step 2 examinations. ChatGPT attained accuracies of 44% and 42% on AMBOSS-Step1 and AMBOSS-Step2, respectively, and 64.4% and 57.8% on NBME-Free-Step1 and NBME-Free-Step2, respectively. This study showcased the capability of ChatGPT in medical education and the evaluation of knowledge. Conducting a trial of a substantial language model (ChatGPT) in general practice alongside the Applied Knowledge Test (AKT) serves as another illustrative example that highlights the possibilities and constraints of AI chatbots in primary care (Thirunavukarasu et al., 2023a). This research scrutinized the merits and demerits of ChatGPT in primary care through the employment of the membership of the Royal College of General Practitioners Applied Knowledge Test (AKT) as a medium. The findings of this study revealed that the average overall performance of ChatGPT amounted to 60.17%, which falls below the average passing threshold of the past 2 years (70.42%). This investigation furnishes evidence that ChatGPT harbors promise in primary care settings, but further research is requisite to enhance its performance. In a separate work, Wang et al. (2023b) performed a case study on the utility of ChatGPT in pharmaceutical development for the creation of a drug to combat addiction to cocaine. In this study, GPT-4 was utilized as a virtual mentor, providing valuable strategic and methodological guidance to researchers focused on developing generative models for potential drug candidates. The main objective of this study was to generate drug-like molecules that possess the desired properties, and by harnessing the capabilities of ChatGPT, this study introduced a fresh and innovative approach to the drug discovery process. So, Chatbots assume the role of facilitators, directing researchers toward inventive methodologies and fruitful avenues for creating efficacious drug candidates.

ChatGPT has passed the theoretical component of the US medical licensing examination despite needing additional training or in-depth medical education (Au Yeung et al., 2023). But employing ChatGPT and other AI technologies in medical writing raises moral and legal concerns, like the need for transparency in AI-generated work and potential copyright breaches. Even though ChatGPT has a lot of promise to improve patient-centered care in radiology and other medical disciplines, more investigation is needed to determine whether it has any disadvantages. Inaccurate statements and possibly dangerous outcomes have all been reported. Therefore, it is essential to consider the limitations of ChatGPT and other foundation models, mainly when used in the medical sector (Wen and Wang, 2023). The usage of ChatGPT as an embodied conversational agent in a social VR platform illustrates the potential of integrating external frameworks into social VR to create new sorts of collaborative experiences (Numan et al., 2023). It demonstrates how adaptable ChatGPT is as a tool for improving user experiences across various scenarios. In the future, ChatGPT will be highly beneficial in mental health. Participants in a study conducted by the mental health software start-up Koko had access to a therapeutic conversational experience where they could talk to ChatGPT about their mental health issues (Grodniewicz and Hohol, 2023). Chatbots can also offer services and information in public health, such as helping smokers quit (Ogilvie et al., 2022). ChatGPT can deliver individualized and accurate medical information, leading to fewer errors and complications, improved patient outcomes, greater patient involvement, and more affordable healthcare delivery (Li et al., 2023). However, to assess the efficacy of ChatGPT in medicine, more research is required because the examined reports need to be of better quality. Despite this, the usage of ChatGPT in the medical industry has a lot of potential. Future research should thoroughly document its functionality, distribution strategies, and theoretical underpinnings (Payton et al., 2017).