Chatbots have been used in healthcare for decades, with the earliest known healthcare chatbot being ELIZA, created in 1966 [26]. Since then, chatbots have been utilized in various healthcare settings, from helping patients manage chronic conditions to providing mental health support. There are three main types of healthcare chatbots based on the input processing and response generation method: rule-based model, retrieval-based model, and generative model [8]. Rule-based chatbots use pre-programmed responses to provide information to patients, while retrieval-based bots offer more flexibility as it queries and analyzes available resources. The generative model produces chatbots based on machine or deep learning to improve their responses over time and is more promising for several reasons. They can understand natural language and learn from user interactions to provide more personalized responses. Additionally, AI-based chatbots can be used to analyze large amounts of patient data to identify patterns and trends that can be used to improve healthcare outcomes [27]. GPT (generative pre-trained transformer) is a well-known representative of such AI-chatbots and has been the subject of many discussions [28].

There are ample studies in the literature describing the potential prospects and applications of chatbots in urology. According to our findings, they are already in the focus of urological symptom checking, health screening, patient education, counseling, lifestyle change, conservative management, clinical decision support, and post-treatment follow-up care. The same was shown by Calvo et al. [29], who conducted a study on the feasibility and usability of a text-based conversational agent that processes a patient’s text responses and short voice recordings to calculate an estimate of their risk for an asthma exacerbation. The chatbot offers follow-up information for lowering risk and improving asthma control, to improve understanding and self-management of the condition. Ferré et al. [30] developed a chatbot-based tool, called the MyRISK score, which collects self-reported patient data before pre-anesthetic consultation to stratify patients according to their risk of postoperative complications. The tool was developed using the Delphi method and logistic regression analysis, with a machine learning model trained to predict the MyRISK score. The tool was found to be effective in predicting postoperative complications, with high sensitivity (94%) but low specificity (49%).

However, this is not a full list of their potential, which is obvious looking at papers investigating AI-chatbots in different medical fields. They can be also valuable for medical education, pre-operative preparation, and academic writing. So, Han et al. [31] developed an AI chatbot educational program to promote nursing skills and found that the experimental group showed significantly higher interest in education and self-directed learning compared to the control group. These studies collectively suggest that chatbots hold promise as a valuable tool for medical education. Chetlen et al. [32] went about the deployment of a chatbot to provide evidence-based answers to frequently asked questions for patients scheduled to undergo a breast biopsy procedure.

By streamlining processes and reducing wait times, chatbots can increase the overall efficiency of the healthcare system, leading to cost savings for healthcare organization [33]. They can also prevent unnecessary office visits and hospitalizations by providing patients with timely and accurate information and support. Chatbots can increase patient engagement and satisfaction by offering personalized advice, information about their condition, and treatment options [34]. Studies have shown that chatbots are more engaging and interactive than traditional online forms, despite taking longer to complete [35•]. According to systematic review from Geoghegan et al. [36], engagement rate for chatbots in the follow-up of patients who have undergone a physical healthcare intervention was up to 97%. In summary, chatbots have the potential to revolutionize the field of urology by improving patient care, optimizing workflow, and increasing the efficiency of the healthcare system.

One of the major challenges facing chatbots in urology is their technical limitations. While chatbots have the potential to improve patient care and physician efficiency, they may not always function as intended due to technical issues. For example, chatbots may experience system failures, errors, or glitches that can affect their performance and accuracy [37]. This can be particularly concerning when it comes to providing medical advice or making diagnoses, as any errors or inaccuracies could lead to serious harm to patients. Therefore, it is essential to thoroughly test chatbots and ensure that they are operating correctly and providing accurate information.

Another challenge that needs to be addressed when using chatbots in urology is privacy and security [38]. Chatbots may store personal health information, which raises concerns about data privacy and security. Patients need to be assured that their personal information is secure and protected from unauthorized access. Furthermore, any data breaches or security incidents could have significant consequences, including loss of patient trust and legal repercussions. Data privacy and security in health chatbots are still under-researched, and related information is underrepresented in scientific literature [39] Therefore, chatbots must comply with relevant privacy and security regulations to safeguard patient data.

Ensuring reliability and accuracy is one of the most crucial factors for the success of chatbots in urology. To provide trustworthy information to patients and healthcare providers, they must be developed with reliable data sources and algorithms [40••]. They should also undergo continuous testing and updates to maintain their reliability and accuracy over time. However, if chatbots are not reliable or accurate, they could misinform patients, leading to incorrect diagnoses or treatments. For instance, Ben-Shabat et al. [41] evaluated the data-gathering function of eight chatbot symptom-checkers and found that the overall recall rate for all symptom-checkers was 0.32 (2280/7112; 95% CI 0.31–0.33) for all pertinent findings. These results suggest that the data-gathering performance of currently available symptom checkers is questionable. As new chatbots become available, hypotheses about their future utility in medicine are limited only by researcher’s imagination. However, their current use should be limited to low-risk tasks with continued human oversight [11]. Regarding to scientific writing, several ethical issues arise about using these tools, such as the risk of plagiarism and inaccuracies, as well as a potential imbalance in its accessibility between high- and low-income countries, if the software becomes paying. For this reason, a consensus on how to regulate the use of chatbots in scientific writing will soon be required [42•].

Chatbots in healthcare face the challenge of resistance from patients and healthcare providers who are unfamiliar with new technology or prefer face-to-face interactions. To overcome this challenge, chatbots need to be designed to be user-friendly and easily integrated into existing workflows. Healthcare providers should also receive training on how to effectively use and recommend chatbots. As Goldenthal et al. [25] indicated, frequent reasons for not activating the chatbot included misplacing instructions for chatbot use, relying on follow-up with clinic or discharge materials, inability to activate chatbot, and inability to text. Moreover, they are not capable of empathy, notably to recognize users’ emotional states and tailoring responses reflecting these emotions. The lack of empathy may therefore compromise the engagement with health chatbots [43].

Our review has some limitations that need to be addressed. Firstly, we focused only on AI-based Chatbots, and incorporating other types of chatbots could expand their clinical application. Secondly, we conducted a qualitative assessment of the literature without collecting all identifiable articles, which could have provided a more comprehensive analysis. Thirdly, some scenarios for the use of AI bots could be further subdivided or combined. However, the purpose of our work was to provide an overview of the prospects and limitations of AI-chatbots in urology. By examining the current literature and exploring various use cases, we hoped to provide an analysis of the potential benefits and drawbacks of their implementing.