Expert perspectives on the use of artificial intelligence in surgical endoscopy: ethical and legal considerations in Türkiye

Authors

Abstract

AimNumerous AI-assisted endoscopic technologies have been developed to overcome the major limitations encountered in endoscopic procedures. However, the introduction of these new technologies has raised various concerns regarding ethical, legal, and accountability-related issues. In this study, we aimed to evaluate the ethical concerns and general perspectives of general surgery specialists and subspecialists in Türkiye regarding the use of AI in endoscopic procedures.
MethodsThis 31-item survey study, which assessed demographic characteristics, ethical concerns, and general perspectives, was conducted across Türkiye between May 1 and June 30, 2025.
ResultsIt was observed that the majority of participants had limited knowledge and a moderate level of concern regarding the use of artificial intelligence (AI) in endoscopic patient management, as well as the associated ethical and legal regulations. Participants emphasized that AI should be used as an assistive tool, provided that appropriate oversight and training are in place. Gastrointestinal surgeons expressed significantly greater concern about potential errors that AI systems might generate, whereas surgical oncologists were more supportive of using AI as an assistive tool. Meanwhile, general surgeons more prominently highlighted the necessity of formal education on AI.
ConclusionGeneral surgery specialists and subspecialists in Türkiye have expressed their ethical and legal concerns regarding the use of artificial intelligence (AI) in endoscopy, along with the perceived need for regulation and education. It is of great importance to consider the perspectives of these specialists during the development and clinical integration of AI systems.

Keywords

Introduction

Artificial intelligence (AI) refers to machine or computer systems that autonomously perform tasks by mimicking human cognitive functions such as learning, problem-solving, and decision-making.¹ Initially developed to assist in diagnostic processes in radiology, AI’s major advantages include its ability to recognize structures and objects with high sensitivity and specificity, provide rapid algorithm-based reports, and deliver highly consistent results.² With rapid technological advancements, AI has attracted substantial attention and has increasingly been integrated into various aspects of human life. Today, AI applications are widely utilized in many fields of medicine, including diagnosis, treatment, drug development, support for patient-physician interactions, and even medical education.^3,4 In recent years, AI has generated considerable interest in healthcare, leading to discussions about whether AI-powered “robot doctors” could eventually replace human physicians in the future.⁵
The integration of AI into endoscopic procedures has emerged as one of the key focus areas in current research, and the advancement of AI applications in medicine continues to evolve through numerous clinical studies in this field.⁶ The main challenges in endoscopic procedures are associated with the complexity of real-time image interpretation and the risk of human error. Some of the most significant limitations in contemporary endoscopic screening and surveillance include interobserver variability in lesion detection, time-consuming biopsy protocols, sampling errors, and the potential to miss subtle or early-stage premalignant lesions.⁷ To address these challenges, AI-assisted endoscopy technologies have been developed, and numerous AI tools have been introduced to enhance lesion detection, characterization, and diagnostic decision-making in endoscopy.^8,9
However, as with all widely adopted technologies, the use of AI in healthcare has raised ethical concerns. Issues such as data privacy and security, lack of transparency, algorithmic bias, accountability, and the mechanization of healthcare services have been subjects of debate.¹⁰ Although there is a growing consensus regarding the potential benefits of AI, ethical and legal concerns remain prominent. This survey-based study aimed to evaluate the ethical concerns and general perspectives regarding the use of AI in diagnosis, patient management, and treatment planning during endoscopic procedures, based on the opinions of actively practicing general surgeons and subspecialists performing endoscopy in Türkiye.

Materials and Methods

This survey-based study investigated the general and ethical perspectives of general surgeons on the use of artificial intelligence (AI) in diagnosis, patient management, and treatment planning during endoscopic procedures. The study was approved by the Non-Interventional Clinical Research Ethics Committee of Samsun University (Date: 30.04.2025, Decision No: GOKAEK 2025/9/14) and was conducted in accordance with the ethical principles outlined in the Declaration of Helsinki. The survey was created using the Google Forms platform and was pre-tested by a group of 10 active general surgeons performing endoscopy who were not involved in the study. The final version of the questionnaire was distributed via email through the Turkish Surgical Society Secretariat and shared on social media platforms between May 1 and June 30, 2025, targeting active general surgeons, surgical oncologists, and gastrointestinal surgeons performing endoscopy across Türkiye.
Considering that approximately 3.500 active general surgeons and surgical subspecialists perform endoscopy in Türkiye, the sample size was calculated to achieve at least 157 participants for a reliable sample with an 80% confidence level and a 5% margin of error. To account for potential data loss, responses from at least 160 participants were targeted. The survey was originally developed and administered in Turkish. An English translation of the survey was provided as supplementary material (Table 1). Informed consent was obtained from all participants, and participant confidentiality and anonymity were maintained throughout the study. Those who did not provide consent were not allowed to access the survey. The questionnaire consisted of 31 questions. The first six questions collected demographic information, including age, gender, years of professional experience (<5 years, 6–10 years, 11–15 years, 16–20 years, 21–25 years, and >25 years), academic title (Specialist, Assistant Professor, Associate Professor, and Professor), type of institution (State Hospital, University Hospital, Private Hospital, and Private Clinic), and surgical specialty or subspecialty (General Surgery, Surgical Oncology, and Gastrointestinal Surgery). Participants were then asked to respond to 25 questions related to three main topics: the use of AI in diagnosis, patient management, and treatment planning in endoscopy, as well as associated ethical concerns. The questions in these main sections included 17 items rated on a five-point Likert scale, while the supplementary section included additional questions using two- to four-point Likert scales (Table 1).
Ethical ApprovalThis study was approved by the Ethics Committee of Samsun University (Date: 30.04.2025, Decision No: GOKAEK-2025/9/14).
Statistical AnalysisDescriptive statistics were used to summarize the characteristics of the study groups. Quantitative variables were presented as mean ± standard deviation (x ± SD), and categorical variables were expressed as counts (n) and percentages (%). Differences between groups were analyzed using chi-square tests. A p-value of less than 0.05 was considered statistically significant. All statistical analyses were performed using a commercial statistical software package (IBM SPSS Statistics for Windows, Version 22.0; IBM Corp., Armonk, NY, USA).
Reporting GuidelinesThis study was reported according to the STROBE guidelines.

Results

A total of 167 participants completed the survey within the specified study period. The female-to-male ratio among the participants was 21/146. The mean age of the participants was 42.12 ± 10.11 years. The majority of participants were general surgeons (n = 134, 80.2%) and those working in state hospitals (n = 90, 53.9%). The descriptive characteristics of the participants are presented in Table 2. A total of 49.7% of participants reported having a moderate level of knowledge about artificial intelligence (AI) applications (Question A1). Additionally, 44.9% stated that they had limited knowledge regarding the use of AI in diagnosis, patient management, and treatment planning during endoscopic procedures (Question A2). Furthermore, 43.7% of the participants indicated that they had no knowledge of the ethical principles and legal regulations concerning AI use either in Türkiye or globally (Question A6). The majority of participants expressed a moderate level of concern regarding the ethical, legal, and data security aspects of AI applications (Question B). Furthermore, 52.7% of participants agreed that AI applications could serve as a supportive tool in physicians’ decision-making processes (Question C1). Most participants strongly agreed that AI applications should be subject to oversight, that formal training should be provided regarding their use, and that detailed information should be communicated to patients (Questions C2–C6) (Figure 1).
Regarding ethical concerns, 73.1% of participants agreed that AI applications could potentially raise ethical issues in certain situations (Question D2), and 77.8% supported the idea that AI should be used solely as an assistive tool (Question D3). Moreover, 47.9% agreed that shared responsibility should apply in cases where AI systems produce erroneous outcomes (Question D4), and 32.3% agreed that AI could serve as a bridge in the physician-patient relationship (Question D6). Additionally, the majority of participants agreed that there is a lack of sufficient legal regulations regarding patient privacy and AI in Türkiye, and expressed the need for education on AI use and ethical principles (Questions D5, D7–D8).
When the survey responses were analyzed according to participants’ gender, years of professional experience, academic title, type of institution, and general surgery subspecialty, no statistically significant differences were observed in any question based on academic title or type of institution. However, statistically significant differences were identified in six questions based on gender, years of professional experience, and general surgery subspecialty (p<0.05). Concerns regarding algorithmic bias and erroneous responses of AI systems were significantly higher among male participants (p=0.007) and gastrointestinal surgeons (p=0.046), with these groups more frequently selecting the “quite concerned” option (Question B2). The rate of participants who had never used AI applications was significantly higher among male surgeons (p=0.038) and those with more than 26 years of professional experience (p=0.009) (Question D1). The belief that AI applications would inevitably cause ethical issues was significantly lower among participants with 6–10 years (2.6%) and 21–25 years (30.8%) of professional experience (p=0.040) (Question D2). Agreement with the statement that AI applications could assist physicians in the decision-making process was significantly higher among surgical oncologists (p=0.006) (Question C1). Notably, gastrointestinal surgeons more frequently selected “strongly disagree” regarding the necessity of comprehensive pre-implementation testing for AI applications (p=0.011); however, across all groups, the “strongly agree” option had the highest proportional response rate (Question C2). Regarding the necessity of providing regular training on AI use, agreement rates were significantly higher among surgical oncologists, while strong agreement was significantly more frequent among general surgeons (p=0.012) (Question C3) (Table 3).

Discussion

In our study, the majority of participants reported having little or no knowledge regarding the use of artificial intelligence (AI) in endoscopic diagnosis, patient management, treatment planning, and the related ethical and legal regulations. Most participants expressed moderate concerns regarding ethical issues, legal liability, and data security associated with AI applications. There was a high level of agreement among participants regarding the use of AI as an assistive tool in clinical decision-making, the need for regulatory oversight, the importance of education on AI use and ethics, and the necessity of informing patients in detail. Male participants demonstrated significantly lower rates of AI use but expressed significantly greater concerns about erroneous outputs of AI systems. Participants with 21–25 years of professional experience were significantly less likely to believe that AI would inevitably cause ethical issues. Concerns about erroneous responses from AI applications were significantly higher among gastrointestinal surgeons, while surgical oncologists strongly agreed on the necessity of AI as an assistive tool, and general surgeons emphasized the need for regular training on AI use.
With advances in medicine, there has been a growing need for innovative technologies and new methods to address diagnostic, treatment, and management gaps in gastrointestinal (GI) diseases. Owing to its strong data analysis capacity and increasing popularity, AI has shown high potential in addressing these challenges.¹¹ Deep learning-based imaging and detection algorithms have been integrated into wearable medical devices for early disease alerts and healthcare management.¹² Consequently, some AI algorithms have begun to enter routine clinical practice in upper and lower GI endoscopy, assisting with early diagnosis, differential diagnosis, and assessment of tumor invasion depth.^13,14,15
However, despite progress in AI applications in endoscopy and other medical fields, issues such as data standardization and sharing, algorithm interpretability and safety, and unresolved ethical and legal concerns remain.¹¹ Particularly in the diagnosis and treatment of GI cancers, the clinical utility of AI applications depends not only on high diagnostic accuracy but also on clinicians’ understanding of and confidence in these technologies.¹⁶ The complexity of AI models, with millions of parameters, their reliance on specific datasets, limited applicability to real-world data, and the lack of continuous learning mechanisms, all hinder transparency and trust in decision-making processes. Moreover, since 2022, only approximately 43% of AI-based medical devices approved by the U.S. Food and Drug Administration (FDA) have published clinical validation data, and less than 5% have been validated through prospective randomized controlled trials.¹⁷ This situation negatively affects clinicians’ trust in and acceptance of AI for diagnostic, therapeutic, and follow-up purposes.^18,19,20 In our study, most participants expressed moderate concerns about data security in AI applications. Male participants and gastrointestinal surgeons were significantly more concerned about erroneous responses from AI systems.
During GI endoscopy, the accuracy of optical diagnosis often falls below expectations and varies significantly among endoscopists. To address this, computer-aided diagnosis (CADx) using AI has been introduced. Recent prospective studies have shown that collaborative use of AI with human endoscopists does not significantly improve diagnostic accuracy compared to human judgment alone.^21,22 Furthermore, one randomized trial demonstrated that fully autonomous AI systems (without human assistance) outperformed human-AI collaboration in diagnostic accuracy.²³ Although autonomous AI tools are not yet widely adopted in endoscopy, many endoscopists perceive them as either too futuristic or legally problematic, emphasizing the need for strict oversight of every aspect of model development to ensure fairness, safety, efficacy, and ethical compliance. Regular audits, algorithm transparency, and clinician training are considered crucial for building trust and accountability.^24,25 In our study, the majority of participants agreed that AI should only serve as an assistive tool in clinical decision-making and that its applications should be subject to regulatory oversight.
Another critical aspect of using AI in clinical endoscopic practice involves ethical, legal, safety, and regulatory factors. In particular, the use of AI in healthcare raises concerns about patient privacy violations, algorithmic bias leading to unfair outcomes, and uncertainties regarding accountability for AI-driven decisions.¹¹ When an AI system provides a misdiagnosis, questions arise regarding whether responsibility lies with the clinician, the hospital, regulatory bodies, or the system developers. Moreover, trust in autonomous systems remains a major barrier, as delegating diagnostic authority to machines and excluding human involvement may cause significant discomfort among patients and healthcare professionals.²⁴ In our study, most participants reported moderate concerns regarding AI applications and indicated limited knowledge about the use of AI in endoscopic diagnosis, patient management, treatment planning, and relevant ethical and legal regulations. They also agreed on the need for education on AI use and ethics, as well as the necessity of detailed patient information. Participants with 21–25 years of professional experience were significantly less likely to believe that AI would inevitably cause ethical issues, and general surgeons strongly agreed on the need for regular training in AI applications.

Limitations

This study has several limitations. First, it was a content-focused survey conducted solely among general surgeons, gastrointestinal surgeons, and surgical oncologists, which may limit the generalizability of the findings to the broader healthcare system. Although the sample size was calculated appropriately, a larger national sample considering variables such as age, gender, years of professional experience, academic title, type of institution, and surgical specialty or subspecialty would have strengthened the study’s results. Furthermore, while the study assessed ethical concerns, larger-scale approaches such as the Delphi method could have been used for a more comprehensive evaluation; however, this would have made it more difficult to reach a sufficiently large sample.

Conclusion

In conclusion, this study revealed that general surgery specialists and subspecialists in Türkiye have limited knowledge regarding the use of AI applications in endoscopy and the related ethical and legal regulations. They also expressed concerns about trust, ethical, and legal issues. There was a consensus that AI should be used solely as an assistive tool in clinical decision-making, that its applications should be regularly audited, and that formal training should be provided. Moving forward, it is essential to incorporate the perspectives of general surgery specialists and subspecialists—who are potential end users—into the development of these systems to comprehensively address issues related to trust, legal and ethical considerations, oversight, and education.

Declarations

Abbreviations

AI: Artificial intelligence
CADx: Computer-aided diagnosis
FDA: Food and Drug Administration
GI: Gastrointestinal
SPSS: Statistical Package for the Social Science

References

1. Chanchlani R, Acharya H. Acute scrotum in children: a retrospective study of cases with review of literature. Cureus. 2023;15(3). doi:10.7759/cureus.36259
   Article PubMed Google Scholar
2. Lee LK, Monuteaux MC, Hudgins JD, et al. Variation in the evaluation of testicular conditions across United States pediatric emergency departments. Am J Emerg Med. 2018;36(2):208-212. doi:10.1016/j.ajem.2017.07.078
   Article PubMed Google Scholar
3. Ogbetere F. Testicular torsion: losses from missed diagnosis and delayed referral despite early presentation. Niger J Clin Pract. 2021;24(5):786-788. doi:10.4103/njcp.njcp_432_20
   Article PubMed Google Scholar
4. Doğan M. Akut skrotal ağrı [Acute scrotal pain]. In: Taşar M, ed. Çocuk Acil Kliniğinde Sık Görülen Başvuru Semptomlarına Yaklaşım. Türkiye Klinikleri; 2022:91-95.
   PubMed Google Scholar
5. Cokkinos DD, Partovi S, Rafailidis V, et al. Role and added value of contrast-enhanced ultrasound of the painful scrotum in the emergency setting. J Ultrasound. 2023;26(2):563-575. doi:10.1007/s40477-022-00764-2
   Article PubMed Google Scholar
6. Islam MM, Satici MO, Eroglu SE. Unraveling the clinical significance and prognostic value of the neutrophil-to-lymphocyte ratio, platelet-to-lymphocyte ratio, systemic immune-inflammation index, systemic inflammation response index, and delta neutrophil index: an extensive literature review. Turk J Emerg Med. 2024;24(1):8-19. doi:10.4103/tjem.tjem_198_23
   Article PubMed Google Scholar
7. Tuşat M, Memiş S. Evaluation of inflammatory markers and HALP score in childhood intussusceptions. Ann Clin Anal Med. 2025;16(7):530-534. doi:10.4328/acam.22763
   Article PubMed Google Scholar
8. Cui P, Cheng T, Yan H. The value of NLR and PLR in the diagnosis of rheumatoid arthritis combined with interstitial lung disease and assessment of treatment effect: a retrospective cohort study. Int J Gen Med. 2025;18:867-880. doi:10.2147/ijgm.s509546
   Article PubMed Google Scholar
9. Xu N, Zhang JX, Zhang JJ, et al. The prognostic value of the neutrophil-to-lymphocyte ratio and platelet-to-lymphocyte ratio in colorectal cancer and colorectal anastomotic leakage patients: a retrospective study. BMC Surg. 2025;25(1):57. doi:10.1186/s12893-024-02708-5
   Article PubMed Google Scholar
10. Pan L, Xiao J, Fan L. The value of NLR, PLR, PCT, and DD levels in assessing the severity of hyperlipidemic acute pancreatitis. Front Med (Lausanne). 2025;12:1561255. doi:10.3389/fmed.2025.1561255
    Article PubMed Google Scholar
11. Zahorec R. Neutrophil-to-lymphocyte ratio, past, present, and future perspectives. Bratisl Lek Listy. 2021;122(7):474-488. doi:10.4149/bll_2021_078
    Article PubMed Google Scholar
12. Serebryanaya N, Shanin S, Fomicheva E. Blood platelets as activators and regulators of inflammatory and immune reactions, part 1: basic characteristics of platelets as inflammatory cells. Med Immunol (Russ). 2018;20(6):785-796. doi:10.15789/1563-0625-2018-6-785-796
    Article PubMed Google Scholar
13. Sağır S. Using NLR, PLR, and ALT*WBC index to differentiate between epididymo-orchitis and testicular torsion in children: a comparative study. Med Pharm J. 2023;2(3):167-174. doi:10.55940/medphar202354
    Article PubMed Google Scholar
14. Arslan S, Azizoglu M, Kamci TO, et al. The value of hematological inflammatory parameters in the differential diagnosis of testicular torsion and epididymo-orchitis in children. Exp Biomed Res. 2023;6(3):212. doi:10.30714/j-ebr.2023.185
    Article PubMed Google Scholar
15. Bayne CE, Villanueva J, Davis TD, et al. Factors associated with delayed presentation and misdiagnosis of testicular torsion: a case-control study. J Pediatr. 2017;186:200-204. doi:10.1016/j.jpeds.2017.03.037
    Article PubMed Google Scholar
16. Vasconcelos-Castro S, Soares-Oliveira M. Abdominal pain in teenagers: beware of testicular torsion. J Pediatr Surg. 2020;55(9):1933-1935. doi:10.1016/j.jpedsurg.2019.08.014
    Article PubMed Google Scholar
17. Dias Filho AC, Rocha RS. Is it time to also use hematologic parameters for the diagnosis and prognosis of testicular torsion? Transl Androl Urol. 2022;11(10):1368.
    PubMed Google Scholar
18. Keseroğlu BB, Güngörer B. Predictive role of large unstained cells and hematological data in the differential diagnosis of orchitis and testicular torsion. Mid Blac Sea J Health Sci. 2021;7(1):97-103.
    PubMed Google Scholar
19. Çoşkun N, Metin M. Can the pan-immune inflammation value be used as a biomarker for differentiating epididymo-orchitis and testicular torsion in children? Medical Records. 2025;7(2):500-506. doi:10.37990/medr.1647261
    Article PubMed Google Scholar
20. Yığman M, Ekenci BY, Durak HM. Predictive value of hematologic parameters in the differential diagnosis of testicular torsion and epididymo-orchitis. Androl Bul. 2024;26(3):167-172. doi:10.1007/s11255-024-04151-0
    Article PubMed Google Scholar
21. Ural DA, Karakaya AE, Güler AG. Comparative analysis of the acute appendicitis management in children before and during the coronavirus disease 2019 pandemic. KSU Med J. 2023;18(1):120-125.
    PubMed Google Scholar
22. Krzywinska E, Stockmann C. Hypoxia, metabolism, and immune cell function. Biomedicines. 2018;6(2):56. doi:10.3390/biomedicines6020056
    Article PubMed Google Scholar
23. Çağlar U, Yıldız O, Ayrancı A. Predictive role of hematological parameters in testicular torsion and epididymo-orchitis in adolescents. Androl Bul. 2025;27(1):72-77. doi:10.24898/tandro.2025.34545
    Article PubMed Google Scholar
24. Lee HY, Lim DG, Chung HS, et al. Mean platelet volume is the most valuable hematologic parameter in differentiating testicular torsion from epididymitis within the golden time. Transl Androl Urol. 2022;11(9):1282. doi:10.21037/tau-21-1112
    Article PubMed Google Scholar
25. Benlioğlu C, Ali Ç. Comparison of hematological markers between testicular torsion and epididymo-orchitis in acute scrotum cases. New J Urol. 2021;16(3):207-214. doi:10.33719/yud.2021;16-3-857388
    Article PubMed Google Scholar

Additional Information

Publisher’s Note
Bayrakol MP remains neutral with regard to jurisdictional and institutional claims.

Rights and Permissions

Creative Commons License

This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License (CC BY-NC 4.0). To view a copy of the license, visit https://creativecommons.org/licenses/by-nc/4.0/

View full license details →

About This Article

Received:

July 16, 2025

Accepted:

August 19, 2025

Published Online:

August 29, 2025

Printed:

September 1, 2025