Translate this page into:
Perceptions of Artificial Intelligence in Healthcare: A Cross-Sectional Need-Based Survey Among Healthcare Workers
*Corresponding author: Nazia Nazir Zargar, Professor, Department of Anaesthesiology, Government Institute of Medical Sciences, Kasna, Greater Noida, Uttar Pradesh, India. nazunazir@gmail.com
-
Received: ,
Accepted: ,
How to cite this article: Mishra S, Singh R, Zargar NN, Khanuja S. Perceptions of Artificial Intelligence in Healthcare: A Cross-Sectional Need-Based Survey Among Healthcare Workers. Glob J Guntur Med Coll. doi: 10.25259/GJGMC_11_2026
Abstract
Objectives:
Artificial intelligence (AI) based tools are rapidly entering clinical workflows. Understanding clinicians’ perceptions, the areas they expect to benefit from AI, desired competency levels, and ethical concerns are essential to plan training and safe implementation.
Material and Methods:
A cross-sectional, anonymized Google survey was distributed to healthcare workers at a medical college. Data were collected between 16 January 2026 and 23 January 2026 (N = 99 respondents). The questionnaire asked about perceived benefits of AI, current clinical challenges that AI could help, workflow areas AI could optimize, why clinicians should understand AI, key ethical concerns, desired competency level, and one measurable outcome that AI training could improve.
Results:
Ninety-nine respondents from a wide range of specialties participated (most frequent specialties: General Surgery 8/99, Microbiology, Pediatrics, General Medicine each 5/99). Mean clinical experience was 10.6 years. The most selected potential benefits were earlier disease detection (74/99, 74.7%) and improved diagnostic accuracy (72/99, 72.7%). AI use in the optimization of clinical notes (85/99, 85.9%) and imaging interpretation (65/99, 65.7%) was documented. Desired AI competency levels were practical use/applying tools safely (46/99, 46.5%), advanced (designing/evaluating models) (40/99, 40.4%), and basic literacy (13/99, 13.1%). The leading ethical concerns were clinician accountability/liability (41/99, 41.4%), patient safety (31/99, 31.3%), transparency (22/99, 22.2%), bias (16/99, 16.2%), and data confidentiality (15/99, 15.2%).
Conclusion:
Clinicians across career stages expect AI to improve detection and diagnostic accuracy and wish for AI training that emphasizes practical, safe use and accountability.
Keywords
Artificial intelligence
Data privacy
Information technology
Perception
Workflow optimization
INTRODUCTION
Artificial intelligence (AI) promises to augment many aspects of healthcare, from image interpretation to predictive risk stratification and administrative automation.1 However, safe and effective adoption depends on clinician trust, competency, and clarity about ethical and medicolegal responsibilities.2 Recently, the National Board of Examinations in Medical Sciences has launched an online course on the use of AI in Healthcare across the country. Hence, we decided to conduct a rapid cross-sectional survey of clinicians at our medical college to capture their expectations, training needs, and ethical concerns about AI in healthcare before the start of the course.
MATERIAL AND METHODS
Study design & setting
Cross-sectional anonymized survey administered via Google Forms to healthcare workers (postgraduate residents, senior residents, faculty, and allied clinicians) affiliated with a medical college. The principal investigator is a trainee at the institution; this survey was an internal needs assessment and educational planning activity.
Data collection period
16 January 2026 to 23 January 2026. Responses (N = 99) were exported from Google Forms and are available as the study dataset. (Appendix 1: AI in healthcare feedback form)
Items on Questionnaire
Demographics (specialty, years clinical experience), multiple-choice items about areas where AI could improve patient care (respondents could select multiple), examples of clinical challenges that AI could address (free text), workflow areas where AI could optimize care, why clinicians (not vendors) should understand AI (free text), the single key ethical concern the respondent wanted to learn more about, desired level of AI competency (basic literacy / practical use / advanced), and one measurable outcome AI training could improve.
Data analysis
Responses were exported and analyzed descriptively. For multiple-choice items, we report counts and percentages. Free-text answers were reviewed and coded thematically to extract common examples and ethical concerns.
RESULTS
A total of 99 healthcare workers completed the form. (AI in healthcare feedback form) mean (range) years clinical experience: 10.6 years (range 1–40).
Respondents represented diverse specialties; the most frequent reported specialties were: General Surgery (14.1%), Anesthesiology (17.1%), Microbiology (8.1%), Pediatrics (13.1%), General Medicine (14.1%), Otolaryngeology (5.1%), Pathology (6.1%), Pharmacology (6%), Radiology (8%), and Pulmonology (8%). The survey form took responses for improvements in patient care, workflow optimization, competency level desired, and career support that AI can provide. [Table 1] The following top patterns were noted.
| Section | Question | Response option | Count / Percentage |
|---|---|---|---|
| 1 | How can AI improve patient care in your specialty? | Improved diagnostic accuracy | 72 (72.7%) |
| Earlier disease detection | 74 (74.7%) | ||
| Personalized treatment planning | 56 (56.6%) | ||
| Reduced medical errors | 64 (64.6%) | ||
| Enhanced clinical decision support | 65 (65.7%) | ||
| 2 | Which areas of your workflow could AI optimize? | Documentation / clinical notes | 85 (85.9%) |
| Imaging interpretation | 65 (65.7%) | ||
| Triage and risk stratification | 44 (44.4%) | ||
| Administrative tasks | 52 (52.5%) | ||
| Population health management | 47 (47.5%) | ||
| 3. | Desired level of AI competency | Basic literacy (understanding concepts) | 13.10% |
| Practical use (applying tools safely) | 46.50% | ||
| Advanced (designing/evaluating models) | 40.40% | ||
| 4 | Why is it important for clinicians to understand AI? | To evaluate bias and fairness | 13.10% |
| To ensure patient safety | 27.30% | ||
| To maintain clinical accountability | 27.30% | ||
| To explain AI-supported decisions to patients | 19.20% | ||
| To meet regulatory and ethical standards | 13.10% | ||
| 5 | How does learning AI support your long-term career? | Keeps skills future-ready | 37.40% |
| Enhances leadership opportunities | 6.00% | ||
| Supports research and innovation | 38.40% | ||
| Improves collaboration with data/IT teams | 18.20% |
AI: Artificial intelligence, IT: Information technology
Top clinical benefit: Earlier disease detection (74.7%)
Top workflow benefit: Documentation / clinical notes (85.9%)
Most desired competency: Practical AI use (46.5%)
Top ethical priorities: Patient safety + Clinical accountability (27.3% each)
Top career motivation: Research and innovation (38.4%) [Figure 1].
Ethical concerns (thematic coding of free text) — number of respondents who explicitly mentioned the topic (one or more keywords):
Clinician accountability/liability: 41/99 (41.4%)
Patient safety: 31/99 (31.3%)
Explainability/transparency (need to explain AI decisions to patients): 22/99 (22.2%)
Bias/fairness: 16/99 (16.2%)
Data privacy/confidentiality: 15/99 (15.2%)
De-skilling / overreliance was rarely invoked explicitly but did appear in free text.
Such concerns are in line with the broader AI discussions. Healthcare professionals are aware of the disruptive potential of AI in medicine. Some medical specialties, particularly those involved with diagnostics like radiology and pathology, may be especially vulnerable.3,4
Representative free-text examples (Verbatim phrasing was coded into themes):
“AI can help predict difficult airway.” (Anesthesia — example of a specific clinical challenge)
“Streamline diagnosis of interstitial lung disease on computed tomography (CT) scan” (Pulmonology — imaging challenge)
“I want to better understand bias and how datasets affect results.” (ethical concern)
“Clinicians must learn AI to explain AI-supported decisions to patients and maintain clinical accountability.” (reason for clinician competency)
DISCUSSION
This single-center needs assessment shows broad clinician interest in the field of AI. It highlights a significant shift in clinical perspectives regarding the integration of artificial intelligence (AI) into medical practice.1 Respondents demonstrate a high level of recognition for AI's potential to enhance clinical outcomes, optimize workflows, and redefine professional development within the healthcare sector [Figure 2]. The responses are consistent with the results of some previous studies where clinicians considered AI to have higher accuracy and efficiency in certain tasks compared to healthcare workers, especially in areas where processing of large data is required.5
High selection of documentation and imaging tasks suggests immediate opportunities for efficiency gains (e.g., automated notes, structured reporting, and AI-assisted image triage) follows similar patterns as documented in previous studies.1,5,6
Clinical impact and patient safety
The primary clinical benefits identified by clinicians were earlier disease detection (74.7%) and improved diagnostic accuracy (72.7%) [Figure 1]. This suggests that AI is viewed as a critical tool for addressing diagnostic delays and errors. This perceived value extends to safety, with 64.6% of respondents noting its potential to reduce medical errors. These findings are closely linked to the perceived necessity of AI literacy for clinicians; 27.3% believe understanding AI is vital for ensuring patient safety, while another 27.3% see it as essential for maintaining clinical accountability [Figure 3]. This underscores a consensus that clinicians must remain the final arbiters of AI-supported decisions to uphold the highest standards of care. These results are consistent with the previously published literature, which has clearly stated that there is increasing use of AI in clinical decision-making and patient care in healthcare systems.6,7
Workflow optimization
Perhaps the most striking finding is the overwhelming demand for AI to assist with documentation and clinical notes (85.9%). This suggests that “administrative burnout” is a major driver of AI interest among clinicians. By automating these non-clinical burdens, AI could potentially return more time to direct patient interactions. Additionally, 65.7% see a strong role for AI in imaging interpretation, indicating a readiness for automated diagnostic support in specialized workflows [Figure 2]. These results are in concordance with previous studies on facilitators and barriers to workflows in healthcare organizations, and their integration with AI.7,8
Educational and professional development
There is a clear appetite for deep technical competency rather than mere awareness. Only 13.1% of respondents are satisfied with basic literacy, while 46.5% desire practical use skills and 40.4% aim for advanced competency in designing or evaluating models. This high demand for advanced training reflects the belief that AI proficiency is crucial for supporting research and innovation (38.4%) and keeping skills future-ready (37.4%) [Figure 3]. Overall, the respondents had a positive outlook toward AI in education, as reported by similar studies on medical students, where prioritizing AI education in the medical curriculum was reported.1,2
Organizational and ethical value
From an organizational perspective, AI-literate clinicians are seen as the key to better technology adoption (46.5%), which suggests that “bottom-up” clinician engagement is vital for the successful implementation of digital tools. Furthermore, clinicians recognize that their role in the AI era includes explaining AI-supported decisions to patients (19.2%) and evaluating bias and fairness (13.1%), indicating that ethical stewardship remains a core component of the medical profession [Figure 4].
These align with international literature on clinician hesitancy: clinicians want to understand how an algorithm arrives at a decision, who is responsible for errors, and whether the tools are fair across patient groups.9-11
The distribution of desired competency levels suggests two simultaneous priorities for educational programs [Figure 5]:
Short-term practical training — how to use available AI tools safely, how to verify outputs, and how to communicate AI-aided decisions to patients and families1 and
Opportunities for advanced training for clinicians interested in designing or evaluating models (researchers and faculty), since ~40% requested advanced competency.11
Implications for implementation
Prioritize clinician-facing, case-based workshops focused on high-value workflows (documentation, imaging, triage) [Figure 6].
Build governance (local validation, monitoring, incident reporting) and clear policies on responsibility and liability.
Require transparency features (explainability) and bias audits before deployment.
Secure data governance: strong de-identification and access controls are necessary to address privacy concerns.
LIMITATIONS
This was a single-center convenience sample survey, with results reflecting the opinions of one medical college only and may not be generalizable. There is a risk of self-selection bias as the clinicians with an interest in AI may have been more likely to respond. The free-text answers were coded post hoc and may not capture all nuances.
CONCLUSION
In summary, the clinicians at the surveyed medical college are optimistic about AI’s potential to improve detection and diagnostic accuracy, and they prioritize practical competency for safe, clinically useful adoption. They see AI not merely as a peripheral technological advancement but as a fundamental tool for improving diagnostic precision and alleviating administrative burdens. To realize these benefits, there is an urgent need for medical education to provide practical and advanced AI training, ensuring that clinicians can lead the ethical and safe implementation of these technologies within their organizations.
Author’s contribution:
NNZ: Survey drafting, circulation, manuscript drafting, data extraction, editing; SM: Data collection, manuscript drafting, data extraction; RS: Manuscript drafting, data extraction, editing; SK: Manuscript drafting, editing.
Ethical approval:
The research/study was approved by the Institutional Review Board at Government Institute of Medical Sciences IECECR/1224/INST/UP/2019, approval number GIMS/IEC/HR/2025/96, dated 28 December 2025.
Declaration of patient consent:
Patient's consent not required as there are no patients in this study.
Conflicts of interest:
There are no conflicts of interest.
Use of artificial intelligence (AI)-assisted technology for manuscript preparation:
The authors confirm that they have used artificial intelligence (AI)-assisted technology solely for language refinement and to improve the clarity of writing. No AI assistance was employed in the generation of scientific content, data analysis or interpretation.
Financial support and sponsorship: Nil.
References
- Knowledge, attitudes, and perceptions of artificial intelligence in healthcare among medical students at Sultan Qaboos University. Cureus. 2023;15:e44887.
- [CrossRef] [PubMed] [Google Scholar]
- Healthcare workers' knowledge and attitudes regarding artificial intelligence adoption in healthcare: A cross-sectional study. Heliyon. 2024;10:e40775.
- [CrossRef] [PubMed] [Google Scholar]
- An artificial intelligence-based mammography screening protocol for breast cancer: Outcome and radiologist workload. Radiology. 2022;304:41-49.
- [CrossRef] [PubMed] [Google Scholar]
- Clinician and computer: A study on patient perceptions of artificial intelligence in skeletal radiography. BMJ Health Care Inform. 2020;27:e100233.
- [CrossRef] [PubMed] [Google Scholar]
- Aspects of artificial intelligence In: Karthikeyan J, Ting SH, Ng YJ, eds. Learning outcomes of classroom research. New Delhi: L'Ordine Nuovo Publication; 2021. p. :48-55.
- [Google Scholar]
- Artificial intelligence and radiology: What will the future hold? J Am Coll Radiol. 2018;15:501-3.
- [CrossRef] [PubMed] [Google Scholar]
- Topol. High-performance medicine: The convergence of human and artificial intelligence. Nat Med. 2019;25:44-56.
- [CrossRef] [PubMed] [Google Scholar]
- Artificial intelligence in healthcare: Transforming the practice of medicine. Future Healthc J. 2021;8:e188-94.
- [CrossRef] [PubMed] [Google Scholar]
- Framing the challenges of artificial intelligence in medicine. BMJ Qual Saf. 2019;28:238-41.
- [CrossRef] [PubMed] [Google Scholar]
- Digital pathology and artificial intelligence in translational medicine and clinical practice. Mod Pathol. 2022;35:23-32.
- [CrossRef] [PubMed] [Google Scholar]
- Attitudes and perception of artificial intelligence in healthcare: A cross-sectional survey among patients. Digit Health. 2022;8:20552076221116772.
- [CrossRef] [PubMed] [Google Scholar]