Awareness, use, and perceived barriers to artificial intelligence in pediatric urology: a multicenter survey

Abstract

Background:

Artificial intelligence (AI) and large language models (LLMs) are increasingly integrated into healthcare, yet their adoption in pediatric urology remains insufficiently explored. Pediatric urology, with its complex and rare conditions, may particularly benefit from AI-based innovations.

Objectives:

This study aimed to assess pediatric urologists’ awareness, usage patterns, and perceptions regarding AI and LLMs, while also identifying potential applications, barriers, and educational needs.

Design:

A cross-sectional, descriptive survey was conducted among pediatric urologists.

Methods:

Between May and July 2025, a 21-item questionnaire was distributed via professional networks and mailing lists. Items addressed demographics, knowledge of AI, frequency and purpose of AI and LLM use, perceived clinical and surgical applications, ethical concerns, and interest in AI training. Descriptive statistics were used for analysis.

Results:

Of 368 invited pediatric urologists, 103 (28%) responded. Most reported moderate (35.0%) or low (29.1%) knowledge of AI, yet more than half (51.5%) used AI tools daily. LLMs had been used by 96.1% of participants, mainly for scientific writing (78.8%), language editing (54.5%), and text summarization (46.5%). Surgical simulation (46.6%), imaging-based strategy planning (40.8%), and preoperative planning (36.9%) were identified as promising clinical applications. Barriers included lack of trust (52.4%), ethical concerns (43.7%), and insufficient knowledge (35.0%). A strong interest in structured AI training was expressed by 81% of participants. Although responses were obtained from multiple countries, the majority of participants were based in Turkey, and the findings should be interpreted accordingly.

Conclusion:

Pediatric urologists demonstrate substantial engagement with AI in academic work, while clinical integration is still limited. The findings highlight a strong demand for AI education and emphasize the need for regulatory clarity, ethical frameworks, and validated tools to enable safe and effective use of AI in pediatric urology.

Plain language summary

How pediatric urologists view and use artificial intelligence in their daily work

Artificial intelligence (AI) is now a common tool in many areas of healthcare, helping doctors with diagnosis, treatment planning, and academic work. In recent years, large language models (LLMs), such as ChatGPT, have also become popular among physicians. However, it is not well known how pediatric urologists, who treat children with urinary and genital problems, use and view these technologies. To explore this, we conducted a survey among pediatric urologists from Turkey and Europe. A total of 103 doctors completed the questionnaire, which asked about their knowledge of AI, how often they use it, what they use it for, and what challenges they face. We found that while most participants reported only a moderate or low level of knowledge about AI, more than half used AI tools on a daily basis. Almost all had tried large language models. These were mainly used for academic purposes such as writing scientific papers, correcting language, and summarizing text. Doctors also saw potential clinical benefits of AI, especially in areas like surgical training, imaging-based planning, and preparing for surgery. At the same time, there were important concerns. Many doctors did not fully trust AI outputs and were worried about who would take responsibility for AI-driven decisions. Ethical issues and data security were also major barriers. Despite these concerns, more than 80 percent of the doctors expressed a strong interest in receiving structured training about AI. In summary, pediatric urologists are already making use of AI in their academic work, but clinical use is still limited. There is a clear demand for education, better regulations, and validated tools to help ensure that AI can be safely and effectively integrated into pediatric urology practice.

Introduction

Artificial intelligence (AI) is increasingly transforming the landscape of healthcare by enabling automation, pattern recognition, and decision support across various specialties. In surgical fields, AI has demonstrated utility in enhancing diagnostic precision, streamlining preoperative planning, and improving postoperative surveillance.^1,2 The growing availability of structured clinical data and computational advances has accelerated the integration of AI tools, including machine learning (ML) and deep learning (DL) algorithms, into clinical workflows. ³

Pediatric urology presents unique challenges due to the diversity of congenital anomalies, age-dependent physiological variations, and the need for long-term follow-up into adolescence and adulthood. Conventional diagnostic and treatment approaches may be limited by subjectivity, variability, and the invasiveness of certain procedures. AI applications, particularly in imaging analysis, anomaly classification, and predictive modeling, offer opportunities to overcome these limitations and support more precise, individualized care for pediatric patients.^1,4 In recent studies, AI tools have shown promising results in areas such as automated grading of vesicoureteral reflux, risk prediction of urinary tract infections, and image-based assessment of hydronephrosis severity.^2,3

In parallel with the expansion of traditional AI applications, the emergence of large language models (LLMs), such as ChatGPT and Gemini, has introduced new possibilities in academic and clinical domains. These models are increasingly used by healthcare professionals to assist with literature review, scientific writing, clinical documentation, and even preliminary decision support. Although not yet validated for direct clinical decision-making, LLMs are widely accessible and user-friendly, prompting their informal integration into the workflows of many physicians and researchers.^5,6 Despite their potential, concerns remain regarding the accuracy, transparency, and ethical implications of relying on such tools, particularly in high-stakes specialties like pediatric surgery.

Although the number of AI-related publications in urology is rapidly increasing, few studies have specifically explored the awareness, usage patterns, and perceptions of pediatric urologists regarding these technologies. Most existing reports focus on technical validation of AI models rather than on their acceptance or perceived clinical utility. Furthermore, the perspectives of pediatric urologists, who manage rare and technically demanding conditions, have not been adequately explored in the current literature. To address this gap, we conducted a descriptive survey to assess pediatric urologists’ familiarity with AI and LLM tools, their current usage habits, and their attitudes toward the ethical, educational, and practical dimensions of AI integration into pediatric urological practice.

Materials and methods

Study design and participants

This study was designed as a cross-sectional descriptive survey conducted between May and July 2025. The target population included physicians actively practicing or training in the field of pediatric urology. Eligible participants were identified through professional pediatric urology networks, national and international society mailing lists, and peer-to-peer distribution. Participation was voluntary and anonymous. Responses were collected either through a web-based platform (Google Forms) or in printed form during institutional distribution. This study was designed and reported in accordance with the Consensus-Based Checklist for Reporting of Survey Studies (CROSS) guidelines. ⁷

The questionnaire was developed based on a review of the current literature on AI integration in clinical practice, with specific emphasis on surgical specialties and urology. The content was refined to reflect areas relevant to pediatric urology, including clinical applications, academic use, and ethical perspectives. The final version consisted of 21 structured items grouped under six domains: (1) participant demographics, (2) AI knowledge and usage patterns, (3) use of LLMs, (4) potential applications of AI in surgical settings, (5) interest in AI-related education and development, and (6) concerns about ethical, professional, or technical barriers. The questionnaire included both multiple-choice and Likert-scale questions to facilitate quantitative analysis and comparability across responses.

A total of 103 pediatric urologists completed the survey out of approximately 368 invited participants, yielding an overall response rate of 28%. The detailed participant recruitment process, including eligibility screening, invitation distribution, and final inclusion, is illustrated in the flow diagram (Figure 1).

Figure 1.

Flow diagram of participant recruitment and inclusion.

The reporting of this study adheres to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines (Supplemental Material). ⁸

Data handling and analysis

All completed responses were included in the final analysis. Data from online submissions and printed forms were combined and reviewed for completeness and consistency. Potential duplicate submissions were checked, but none were identified. Descriptive statistics were used to summarize participant characteristics and responses, with categorical variables presented as frequencies and percentages. No inferential statistics were applied, as the primary aim was to provide a descriptive overview of current practices and attitudes. All analyses were performed using IBM SPSS Statistics version 25.0 (IBM Corp., Armonk, NY, USA).

Ethical considerations

The study protocol was reviewed and approved by the university ethics committee, with approval number E-60116787-020-689602. An introductory statement at the beginning of the questionnaire informed participants about the voluntary nature of participation, ensured anonymity, and stated that no identifiable personal or institutional data would be collected. The study was conducted in accordance with the principles of the Declaration of Helsinki.

Results

Participant characteristics

A total of 103 pediatric urologists participated in the study. All participants completed the full questionnaire and were included in the final analysis. The mean age of the respondents was 38.4 years (SD = 5.6), with a minimum of 32 and a maximum of 59 years. Of these, 64 (62.1%) were practicing in Turkey and 39 (37.9%) in various European countries.

Of the 103 participants, 57 (55.3%) identified as male and 46 (44.7%) as female. The professional titles of the respondents were distributed as follows: 34 specialists (33.0%), 21 assistant professors (20.4%), 20 residents or fellows (19.4%), 16 associate professors (15.5%), and 12 professors (11.7%).

Regarding institutional affiliation, 47 respondents (45.6%) were working at university hospitals, 32 (31.1%) at state hospitals, 20 (19.4%) at private hospitals, and 4 (3.9%) in private outpatient practice settings.

The mean duration of professional experience was 9.5 years (SD = 5.8), ranging from 2 to 29 years. A detailed summary of participants’ demographic characteristics is presented in Table 1.

Table 1.

Demographic characteristics of participants.

Variable	n (%)/Mean ± SD
Number of participants	103
Age (years)	38.4 ± 5.6 (range: 32–59)
Gender
Male	57 (55.3%)
Female	46 (44.7%)
Professional title
Specialist	34 (33.0%)
Assistant professor	21 (20.4%)
Resident/fellow	20 (19.4%)
Associate professor	16 (15.5%)
Professor	12 (11.7%)
Institution
University hospital	47 (45.6%)
State hospital	32 (31.1%)
Private hospital	20 (19.4%)
Private outpatient practice	4 (3.9%)
Years of professional experience	9.5 ± 5.8 (range: 2–29)

Knowledge and use of AI

Participants were asked to rate their self-perceived level of knowledge regarding AI. The most commonly selected response was “moderate” (n = 36, 35.0%), followed by “low” (n = 30, 29.1%) and “high” (n = 22, 21.4%). A smaller number of participants reported “very low” (n = 9, 8.7%) or “very high” (n = 6, 5.8%) levels of knowledge.

When asked about their frequency of using AI or AI-based applications in clinical or academic settings, more than half of the participants (n = 53, 51.5%) reported daily usage. An additional 33.0% (n = 34) stated they used such tools several times per week, while 11.7% (n = 12) indicated monthly use. Only a small proportion of respondents (n = 4, 3.9%) reported never having used AI in their professional activities.

Regarding the use of LLMs such as ChatGPT or Gemini, 99 participants (96.1%) stated that they had experience with these tools. Among these users, 68 (68.7%) reported using the paid version, while 31 (31.3%) used the free version.

Purpose of use

Among the 99 participants who reported using LLMs, 78 (78.8%) stated that they had used these tools for scientific research and writing. Language editing was reported by 54 participants (54.5%), while 46 (46.5%) indicated use for summarization and structuring of text. Similarly, 46 participants (46.5%) reported using LLMs for translation or editing purposes.

Personal correspondence was reported by 47 respondents (47.5%), and 28 (28.3%) had used LLMs for data analysis. Use of LLMs for statistical planning was reported by 24 participants (24.2%), while 34 (34.3%) had employed them for methodology design. Figure or graph creation was indicated by 41 respondents (41.4%).

Less frequent purposes included manuscript writing (n = 20, 20.2%), citation generation (n = 15, 15.2%), new text generation (n = 15, 15.2%), and reference creation (n = 14, 14.1%). A total of 34 participants (34.3%) had used LLMs during the literature search stage of research.

In terms of clinical applications, patient communication was reported by 14 respondents (14.1%), clinical decision support by 10 (10.1%), treatment decisions by 16 (16.2%), medical report writing by 9 (9.1%), and image analysis by 9 (9.1%).

When asked whether they verify the accuracy of AI-generated outputs, 90 participants (90.9%) responded affirmatively. Nine respondents (9.1%) reported that they do not routinely verify AI-generated content.

Perceived applications in surgery

When asked about potential areas in which AI could be beneficial in pediatric surgical practice, 38 participants (36.9%) indicated preoperative planning, while 42 (40.8%) identified the determination of surgical strategy through imaging. A total of 43 respondents (41.7%) selected anatomical evaluation using 3D modeling as a potential application.

Nine participants (8.7%) believed that AI could contribute to intraoperative decision support, and 21 (20.4%) saw value in postoperative complication prediction. Surgical simulation and training were reported by 48 participants (46.6%), and robot-assisted surgery by 45 (43.7%).

Interest in AI training and development

Regarding participation in AI-related training or development processes, 54 participants (52.4%) responded “definitely yes,” and 29 (28.2%) responded “yes.” Twelve participants (11.7%) stated that they were unsure, while 2 (1.9%) did not wish to participate. Six participants (5.8%) reported that they would “definitely not” be interested in participating in such programs.

Ethical concerns and barriers

Participants were asked to identify potential barriers and ethical concerns regarding the implementation of AI in pediatric urology. Among the 103 respondents, the most commonly cited barrier was lack of trust, reported by 54 participants (52.4%). This was followed by responsibility for decisions, selected by 46 participants (44.7%), and ethical issues, also noted by 45 participants (43.7%).

Data security was considered a barrier by 44 participants (42.7%), and lack of knowledge was cited by 36 participants (35.0%). Loss of creativity was reported by 32 participants (31.1%), while overdependence on AI was selected by 21 participants (20.4%). Lack of time to engage with or implement AI tools was the least frequently mentioned concern, reported by only nine participants (8.7%). A summary of these ethical and practical concerns is illustrated in Figure 2.

Figure 2.

Reported ethical and practical concerns regarding AI use in pediatric urology.

When asked about the expected impact of AI on the physician–patient relationship, 48 participants (46.6%) indicated that they anticipated no effect, 27 participants (26.2%) expected a minor improvement, and 12 participants (11.7%) believed there would be a major improvement. On the other hand, 10 participants (9.7%) predicted a minor deterioration, and 6 participants (5.8%) anticipated a major deterioration.

Participants were also asked whether they were concerned that AI could make their profession obsolete in the future. Thirty-seven respondents (35.9%) answered “yes,” while 66 participants (64.1%) reported no such concern.

Discussion

This study provides a comprehensive overview of pediatric urologists’ awareness, utilization patterns, and attitudes toward AI and LLMs. While the use of AI in various medical and surgical specialties has been increasingly studied, research specifically exploring perspectives within pediatric urology remains scarce. Our findings indicate that although these tools, particularly LLMs, are widely used for academic purposes such as manuscript writing, literature synthesis, and presentation development, their integration into clinical workflows is still limited. Nevertheless, participants expressed notable interest in the future potential of AI in areas such as surgical simulation, preoperative planning, and decision support. Ethical and legal concerns, including data privacy, responsibility attribution, and the reliability of AI-generated output, remain significant barriers to widespread implementation.

AI is now widely used across specific areas of medicine. A generally positive attitude toward AI has been reported in these fields. For example, 82% of Norwegian breast radiologists expressed support for integrating AI into the interpretation process of mammographic screening, largely due to its perceived potential to reduce workload and enhance efficiency. ⁹ Similarly, anesthesiologists and intensive care specialists across Europe have demonstrated an openness to AI applications and reported expectations of various clinical benefits associated with its use. ¹⁰

A growing body of survey-based research highlights the increasing optimism among physicians toward the integration of AI in clinical practice. A comprehensive survey conducted among Portuguese medical doctors revealed a general sense of optimism within the medical community and a strong willingness to adopt AI Technologies. ⁶ Similarly, German surgeons have expressed favorable attitudes toward the implementation of AI in surgical settings. ¹¹ In South Korea, the majority of physicians reported believing in the utility of AI in medicine, particularly emphasizing its potential to assist in diagnostic processes and treatment planning. ¹² Moreover, a study among family physicians supported the view that AI could play a supportive role in clinical assessment and diagnostic decision-making. ¹³

Pediatric urology includes rare and heterogeneous congenital conditions that require highly specialized interpretation of imaging. Diagnostic processes such as vesicoureteral reflux grading and hydronephrosis assessment often depend on subjective evaluation and show considerable interobserver variability. AI models can standardize these assessments, reduce diagnostic inconsistency, and improve early risk stratification through objective image analysis. Because clinical experience may vary and high-quality evidence is limited in several pediatric conditions, AI has the potential to enhance precision, reproducibility, and diagnostic confidence in this subspecialty.

In our study, pediatric urologists also demonstrated a high level of interest in AI technologies, with 51.5% of respondents reporting near-daily use of AI tools in their professional activities. These findings indicate that awareness and utilization of AI are extending even into niche specialties such as pediatric urology. Recent publications in this field, including studies on automated vesicoureteral reflux grading, ultrasound-based hydronephrosis evaluation, AI-assisted detection of detrusor overactivity in urodynamic studies, and the reduction of subjectivity in hypospadias classification, further support the growing interest and ongoing integration of AI technologies in pediatric urology.^2,4

Recent consensus guidelines on pediatric urolithiasis also highlight substantial variability in diagnostic practices, the need for pediatric-specific tools, and the overall scarcity of high-quality evidence, which together limit the integration of advanced technologies. Furthermore, these guidelines identify informatics and AI as priority areas for future research because of their potential to reduce clinical variability and support individualized decision-making. ¹⁴ These points align with our findings, which show strong interest among pediatric urologists but persistent concerns regarding reliability, safety, and the lack of validated pediatric-focused AI solutions.

Findings from the recent systematic review on AI in urolithiasis demonstrate consistently high diagnostic and predictive performance across multiple clinical domains, which helps explain the growing interest among pediatric urologists in AI applications. ¹⁵ At the same time, the review highlights the limited multicenter validation and lack of standardized frameworks, aligning with our respondents’ concerns about reliability, safety, and the absence of pediatric-specific AI tools.

The integration of AI, particularly LLMs such as ChatGPT, into scientific manuscript preparation has expanded rapidly, offering notable advantages alongside important limitations. In our study, most pediatric urologists who reported using LLMs did so mainly for language editing (54.5%) and text summarization or restructuring (46.5%). Given these limitations, it is noteworthy that 90.9% of participants routinely verified the accuracy and reliability of AI-generated content. These findings suggest a cautious yet pragmatic approach to AI adoption among clinicians, recognizing its utility in academic work while remaining aware of its potential risks. ¹⁶

AI-supported approaches facilitate personalized treatment planning by integrating patient-specific data and tailoring surgical strategies to individual needs. In pediatric urology, such tools may assist in selecting procedures or grafts, potentially improving outcomes and satisfaction. ¹

AI can also enhance preoperative decision-making by analyzing clinical data and supporting risk assessment. A recent survey among German surgeons showed strong support for AI in surgical planning, particularly when used under expert supervision. ¹¹

Intraoperatively, AI may contribute to decision-making and precision, especially in complex or robot-assisted procedures. In addition, AI-powered surgical simulations are increasingly used in training, offering safe and repeatable practice environments. In our study, 36.9% of pediatric urologists identified preoperative planning as a key application area, while 40.8% highlighted AI’s role in determining surgical strategies. Moreover, 46.6% saw value in AI-based surgical simulation and training, reflecting growing interest in AI across all phases of surgical care.

Despite the growing integration of AI into healthcare, a number of ethical concerns and implementation challenges continue to limit its widespread adoption. Key issues include data privacy, the lack of clear accountability in AI-driven clinical decisions, and concerns about the reliability of AI outputs.^9,13 Clinicians have expressed apprehension over AI’s potential to generate factual inaccuracies (“hallucinations”), introduce bias into decision-making, and weaken the physician-patient relationship.^11,13 Furthermore, the absence of robust medico-legal frameworks, especially in high-stakes areas like diagnosis and treatment planning, contributes to professional hesitation.

In addition to ethical concerns, several practical barriers hinder AI adoption in clinical settings. These include limited clinician education on AI technologies, insufficient technical infrastructure, lack of validated clinical evidence, and fears of professional displacement. ¹¹ Interdisciplinary collaboration between clinicians and developers, along with targeted training and regulatory clarity, is essential to overcoming these obstacles and promoting the responsible use of AI in medicine.

In our study, the most frequently reported barrier to AI adoption among pediatric urologists was a lack of trust in AI-generated outputs (52.4%). Responsibility for decisions (44.7%) and ethical concerns (43.7%) were also commonly cited, followed by insufficient knowledge or training in AI tools (35.0%). These findings mirror those reported in the broader medical literature and emphasize the need for targeted education, transparent regulatory frameworks, and validation of AI tools before they can be fully integrated into pediatric surgical practice.

The substantial demand for AI training observed in our study, where over 80% of respondents expressed interest in structured educational programs, highlights a critical need for the integration of AI-related content into both undergraduate and postgraduate medical curricula. As AI technologies become increasingly embedded in clinical and academic practice, equipping future pediatric urologists with foundational knowledge in data science, algorithmic thinking, and ethical considerations will be essential. Incorporating AI modules into continuing medical education programs could also facilitate safe, effective, and informed adoption across varying levels of clinical expertise.

This study is among the first to specifically explore AI awareness, usage patterns, and perceived barriers among pediatric urologists, a highly specialized and underrepresented group in AI-focused healthcare research. The multinational scope, inclusion of both academic and non-academic practitioners, and diverse institutional representation provide a well-rounded understanding of the current landscape. The comprehensive questionnaire, developed through expert consensus, allowed for nuanced exploration of clinical, academic, and ethical dimensions of AI use.

Limitations

This study has several limitations. First, although the survey included participants from multiple countries, the majority of respondents were based in Turkey, which may limit the generalizability of the findings. Second, as a self-reported survey, responses may be subject to recall and social desirability bias. Third, the questionnaire used in this study was not formally validated, which may affect the accuracy and reliability of the results. Despite efforts to ensure diverse participation, the overall sample size was limited and may not fully represent the global pediatric urology community.

Conclusion

This study offers novel insights into the awareness, utilization, and perceived challenges of AI among pediatric urologists. Despite being a highly specialized field, there is a clear and growing engagement with AI tools, particularly for academic purposes. Participants expressed strong interest in further AI training, while also highlighting key ethical and practical concerns that must be addressed. These findings underscore the importance of integrating AI education into medical curricula and fostering interdisciplinary collaboration to support the responsible and effective implementation of AI in pediatric urology.