Attitude, readiness and utilisation of artificial intelligence: a cross-sectional study among undergraduate medical students in a medical college, Kolkata

INTRODUCTION

Artificial intelligence (AI) is an emerging promising field which is taking over the entire healthcare like a storm globally. AI focuses on creating computer systems capable of human-like cognitive functions, encompassing tasks like image and speech understanding, decision-making, and language translation. In healthcare, AI mainly refers to algorithms and software, for analyzing digital health data for diagnoses, management and follow-up of patients.¹ AI in medicine is an umbrella term that uses algorithms and software that analyze data and digital information to make diagnoses and suggest therapies.² Incorporating AI in the healthcare sector can potentially enhance precision, efficiency, and patient outcomes. The various AI applications in the field of medicine are scanning health data, laboratory reports, radiological imaging, developing protocols for treatment, surgical interventions, drugs and vaccines development, genetic studies, and many more.^3,4 AI applications are also influencing medical education, serving students, educators, and administrators to enhance learning and teaching. Medical students, as future healthcare professionals, are crucial in embracing and utilizing AI in patient care.⁵ Machine learning (ML) and deep learning (DL), subsets of AI, are increasingly relevant in undergraduate and postgraduate medical training.

Despite the potential advantages, the medical community has expressed concerns about AI adoption. A recent debate centered on the possibility of AI algorithms surpassing radiologists, potentially rendering their expertise redundant.⁶ This lack of comprehensive information regarding AI’s impact may deter students from pursuing certain medical specialties. A UK-based study indicated that 49% of respondents were less inclined to choose radiology as a career due to AI. These anxieties may shape undergraduate medical students’ perceptions of AI, despite their limited exposure to it within the undergraduate medical curriculum.⁷ Medical institutions already face challenges with continuously evolving curricula and the demands to incorporate new subjects, alongside a shortage of specialized faculty and appropriate teaching methodologies.

Notwithstanding existing survey data, the attitude of undergraduate medical students towards AI in medicine and their openness to its inclusion in the medical curriculum remains uncertain, particularly within our state and India. With this background in mind, the present study was conducted to assess the attitude towards AI among undergraduate medical students of a medical college in Kolkata and to identify the readiness towards AI among the study participants. This study aimed to assess the utilization of AI among the study participants.

METHODS

A descriptive, observational study with a cross-sectional study design was undertaken among undergraduate medical students in phases I, II, and III (part I and part II) at Institute of Post Graduate Medical Education and Research (IPGME and R), Kolkata from August to October 2024. The study participants who were unwilling to give informed consent for participating in the study were excluded. A Microsoft form questionnaire was prepared and distributed to the students through emails and social media platforms (WhatsApp). Informed consent was obtained from all the eligible study participants via electronic means (e-consent) where the purpose of the study was thoroughly explained. Ethical approval was obtained from the institutional ethics committee (IPGMER/IEC/2024/0609 dated 17.08.24) of IPGME and R. The study contained no names and their emails were masked so that the participants could not be tracked, to maintain anonymity and confidentiality.

This study tool divided into 3 segments to assess AI readiness. The first section collected data across the domains of sociodemographic information, (participants’ age, gender, current medical year, etc.) and their awareness of various AI applications and their usage.

The second section contained items taken from the medical artificial intelligence readiness scale for medical students questionnaire (MAIRS-MS).⁸ The MAIRS-MS questionnaire is a validated study tool, which is designed to assess medical students’ readiness and capability to embrace AI across four distinct domains, namely, cognition (8 questions), ability (8 questions), vision (3 questions), and ethics (3 questions), where each question was assigned with a maximum score of 5 (“strongly agree”) and a minimum score of 1(“strongly disagree”). Attitude towards AI was assessed by the attitude scale developed by Sit et al, which is a 5-point Likert scale, having a total of 11 items for the attitude domain, with options for each item having scores of 5 (‘strongly agree’) to 1 (‘strongly disagree’).

Data collection period lasted for 2 weeks, where Microsoft forms were circulated among the eligible study participants across all phases of MBBS through social media platform. Relevant information was also shared with each batch of students in their social network group. After obtaining written e-consent, the participants were asked to fill out the online questionnaire. Reminders were sent periodically to the students to encourage their participation in the study.

The collected data underwent a thorough verification process for completeness and accuracy followed by tabulation and statistical analyses using IBM SPSS Statistics, version 25.0 (IBM Corporation, Armonk, NY, USA). Descriptive statistics, like frequency and percentage, were calculated for the demographic variables and all items in the questionnaire. The attitude was categorized as ‘favourable’ and ‘unfavourable’ based on the median of the overall attitude score. Score ≥ median was classified as ‘favourable’ while score < median was considered as ‘unfavourable’. Item number 3 in the attitude domain of the scale was reversely scored. Mean (±SD) scores for each of the four sub-domains under the readiness domain of MAIRS-MS were calculated, and an overall mean±SD score was obtained by summing up the mean scores of the four sub-domains. Statistical associations between categorical variables were assessed using the Pearson’s Chi-square test, and a p value <0.05 at 95% confidence interval (CI) was taken as statistically significant.

RESULTS

Sociodemographic characteristics

Nearly 57% of the study participants belonged to the age group of 18-21 years. 74.26% of the study participants were males and 32.27% of the study participants belonged to phase I of the current academic year. 67.49% of fathers and 53.95% of mothers of the participants had completed their graduation. 61.62% currently reside in the institutional hostel.

Attitude towards AI

Most common source of information on AI for the study participants were social media (378), followed by friends (238). 44.24% of the study participants have agreed that AI will play an important role in healthcare, while 37.69% of them agreed that AI may replace some medical specialties in near future. 37.69% agreed that all undergraduate medical students should be taught regarding basics of AI and its various emerging applications in healthcare (Table 1). Almost 56% had a ‘favourable attitude’ towards AI (Figure 1).

Readiness on cognitive factors towards AI

24.61% of study participants disagreed with explaining the basic concepts of data science, while 36.12% agreed to have a basic conception of statistics. Around 28.00% of the participants disagreed with explaining how AI systems are trained, 26.18% of them disagreed with giving definitions on AI basics and its terminology, while and 22.57% disagreed with facing problems while analyzing the data obtained by AI in healthcare. Around one-fourth (25.0%) disagreed with differentiating the functions and features of AI-related tools and applications, while 29.35% agreed with expressing the importance of data collection, analysis, evaluation, and safety for the development of AI.

Readiness on ability factor towards AI

Regarding ability of the students concerning AI, approximately 30% agreed to use AI applications for healthcare problems, 31.37% were willing to explain AI applications in healthcare to patients, and 37.69% found value in using AI for education, service, and research. Additionally, 37.92% agreed they could access, evaluate, use, share, and create knowledge using information and communication technologies, 41.08% agreed to use AI applications based on their purpose.

Readiness on vision factor towards AI

41.76% of participants agreed they could foresee both the opportunities and threats AI technology might present, 42.21% agreed they could explain the strengths and limitations of AI, and 34.76% specifically agreed they could explain the limitations.

Readiness on ethics factor towards AI

Almost 40% participants agreed to adhere to the legal regulations in AI use within healthcare, 41.08% agreed to act ethically when using AI, and 39.27% agreed to handle health data according to legal and ethical standards while using AI.

The overall mean and SD score of readiness toward AI was 66.13±17.45. The ability factor domain was found to have the highest mean±SD score (25.70±8.09) (Table 2).

The differences in mean scores of the domains under ‘readiness’ (cognitive, ability, visual and ethics) across the socio-demographic characteristics of study participants were calculated respectively using ANOVA. It was revealed that the gender of participants, permanent residence and mothers’ education had statistically significant differences with the cognitive domain; phase of MBBS, permanent residence, fathers’ and mothers’ education and awareness on AI had statistically significant differences with the both the ability and vision domain, while phase of MBBS, fathers’ and mothers’ education and awareness on AI had statistically significant differences in mean scores with the ethics domain (Table 3).

Spearman’s correlation revealed a moderately positive correlation between the overall readiness score and attitude score, which was statistically significant (Spearman’s rho- 0.483, p<0.001) (Figure 2).

Utilization of AI

Among the study participants, 292 (66.0%) use an AI-based application, out of which, 85.0% of the study participants use Chat GPT. Among 96.9% of the study participants, studying is the most common purpose for using an AI-based application. Almost half of them (50.0%) use them weekly, while among the daily users (20.0%), 17.1% use them for more than 3 days a week. Almost 35 (12%) faced difficulties while using them, the most common being difficulty in understanding and interpreting the data [11 (31.4%)] (Table 4).

*Multiple response.

Most of the study participants used AI-based applications for the study (283) followed by entertainment (187) (Figure 3).

The majority of participants using AI-based applications did so for studying (283), followed by entertainment (187). Among the 34 participants who reported difficulties, the inability to interpret results (11) and encountering wrong information (9) were the most common problems.

DISCUSSION

This study investigated the attitudes and readiness towards artificial intelligence (AI) among 443 medical students at a government medical college, utilizing a validated online questionnaire. Prior research by Li et al indicated that postgraduate medical students demonstrated greater awareness of medical AI compared to undergraduates.¹⁰ Their study also found a positive correlation between the intention to use medical AI and factors like performance expectancy, habit, enjoyment, and trust. In the present study, a significant portion of participants showed some hesitancy regarding fundamental AI concepts, with 24.61% disagreeing about explaining basic data science concepts and 26.18% disagreeing about defining core AI terminology. Additionally, 22.57% disagreed with the proper analysis of AI-derived healthcare data. However, approximately 56% of the students held a ‘favorable attitude’ towards AI. This underscores the importance of future medical education focusing on enhancing students’ practical skills and designing engaging, easily understandable courses to prepare them for their medical careers in an AI-integrated landscape.

A study by Jebreen et al revealed that a majority of their participants (76.79%) had not received formal AI education before or during their medical studies.¹¹ Nevertheless, about two-thirds believed AI would become commonplace (67.9%) and revolutionize medical fields (68.7%). These participants also reported a lack of prior AI training in their formal medical education (74.5%), highlighting a crucial need to incorporate AI education into medical curricula. Contrasting this, the most frequent sources of AI information for students in the current study were social media (378) and friends (238). A substantial 44.24% of our participants agreed that AI would play a significant role in healthcare, and 37.69% believed AI teaching should be mandatory for all undergraduate medical students. Consistent with Jebreen et al findings, roughly 56% in our study also expressed a ‘favorable attitude’ towards AI in healthcare. These studies collectively suggest that current learning opportunities for AI in medicine are insufficient, emphasizing the necessity for further exploration into the integration of AI in medical practice.

Findings from Pucchio et al indicated that their respondents generally agreed that AI applications would become common and improve medicine.¹² Furthermore, 73% believed they would need to use and understand AI in their careers, and 67% advocated for formal AI instruction in medical education. Similar to our study, Pucchio et al participants reported social media (378 in our study) and friends (238 in our study) as their primary sources of AI information. In our study, 44.24% agreed on AI’s important role in healthcare, and 37.69% thought AI would replace some specialties within their lifetime. A matching 37.69% supported AI teaching for all undergraduate medical students. These findings reiterate the urgent need to integrate AI and its applications into the medical curriculum to encourage its responsible application in healthcare delivery.

Jackson et al study on undergraduate medical students found that only 4.2% believed AI could enhance medical decision accuracy, while 49% saw its potential to improve healthcare accessibility.¹³ Students expressed concerns about the impact on patient-physician relationships (54.5%) and breaches of confidentiality (53.5%). A strong demand for structured AI training, particularly on reducing medical errors (76.9%) and ethical issues (79.4%), was evident. Mirroring our study, Jackson et al participants also cited social media (378 in our study) and friends (238 in our study) as their main AI information sources. In our study, 44.24% agreed on AI’s significant role in healthcare, 37.69% anticipated AI replacing some specialties, and 37.69% supported mandatory AI teaching for undergraduates. Additionally, 41.76% of our participants agreed they could foresee AI's opportunities and threats. These results again point towards the need for more AI-focused workshops and seminars for undergraduate medical students to improve their understanding and utilization of AI in healthcare.

A study by Allam et al found that most participants had an inadequate understanding (76.4%) of AI’s importance and use in medicine, with a majority (87.4%) holding a negative attitude.¹⁴ Their study also reported that around 40% of participants agreed to follow legal regulations and ethical principles in AI use and data handling. In contrast, our current study showed that approximately 28.00% of undergraduate medical students disagreed with understanding the efficient use of AI systems, 26.18% disagreed with defining basic AI concepts, and 22.57% disagreed with properly analyzing AI-derived healthcare data. Despite these reservations, about 56% of our students reported a ‘favorable attitude’ towards AI.

Lugito NPH et al study reported mean scores in cognitive, ability, vision, and ethics domains as 24.52±5.26 (out of 40), 27.78±4.65 (out of 40), 10.57±2.07 (out of 15), and 10.47±2.00 (out of 15), respectively.¹⁵ In contrast, our study found an overall mean readiness score towards AI of 66.13±17.45, with the ability factor showing the highest mean score (25.70±8.09).

Similarly, Tung et al study found that 71% of students believed AI teaching would benefit their careers, and 69.44% supported mandatory AI education.¹⁶ On the MAIRS-MS scale, their students had mean scores of 21/40 (cognitive), 25/40 (ability), 10/15 (vision), and 11/15 (ethics), with an overall mean of 67±14.8 out of 110. Our study, however, showed a comparable overall mean readiness score of 66.13±17.45 on the MAIRS-MS scale, with the ability factor again exhibiting the highest mean score (25.70±8.09).

Hamad et al research also using the MAIRS-MS scale reported the highest mean in the ability domain (23.38±7.16), followed by cognitive, ethics, and vision.⁹ Their overall mean AI readiness score was 64.26 out of 110. Consistent with these findings, our present study also found the highest mean score in the ability factor (25.70±8.09), with an overall mean readiness score of 66.13±17.45.

There is paucity of sufficient relevant literature on attitude, readiness, and utilization among undergraduate medical students of West Bengal regarding AI tools and its various applications in healthcare. A large sample size was considered and done on all four phases of undergraduate medical students. It also addressed quite unexplored issues which helps to resolve the problem in the near future. A possible limitation is the sampling technique using social media limits generalizability.

CONCLUSION

In this study, approximately half of the participants held a positive view regarding AI’s potential in healthcare. A significant majority, around 60%, reported understanding fundamental concepts of both data science and AI and expressed a readiness to utilize AI applications in healthcare settings. This willingness to adopt AI was observed even though these students acknowledged a perceived lack of cognitive skills in this domain. Furthermore, the study identified a moderately positive correlation between the participants’ overall readiness to adopt AI and their general attitude towards it. The primary use of AI-based applications among the students was for academic purposes, with Chat GPT being the most popular tool. However, a small fraction of the participants reported encountering challenges while using these This study sheds key insight into the willingness and outlook towards AI literacy among medical students, emphasizing the necessity for AI-related curriculum in undergraduate medical education.

ACKNOWLEDGEMENTS

The authors would like to express their sincere gratitude to the head of the institutions (principal and dean of students’ affairs) of IPGMER for their kind permission to conduct the study. They also like to say thanks to all the students for their all hands of help and kind cooperation.

Funding: No funding sources

Conflict of interest: None declared

Ethical approval: The study was approved by the Institutional Ethics Committee (IPGMER/IEC/2024/0609 dt. 17.08.24)

REFERENCES

1. 1. Soanes C, Hawker S, Elliott L, eds. Pocket Oxford English Dictionary. 10th edn. New Delhi. Oxford University Press; 2005;44.
2. 2. Briganti G, Le Moine O. Artificial intelligence in medicine: today and tomorrow. Front Med. 2020;7:27.
3. 3. Waymel Q, Badr S, Demondion X, Cotton A, Jacques T. Impact of the rise of artificial intelligence in radiology: What do radiologists think? Diagn Interven Imag. 2019;100(6):327-36.
4. 4. Yun D, Xiang Y, Liu Z, Lin D, Zhao L, Guo C, et al. Attitudes towards medical artificial intelligence talent cultivation: an online survey study. Ann Transl Med. 2020;8(11):708.
5. 5. Chan KS, Zary N. Applications and challenges of implementing artificial intelligence in medical education: integrative review. JMIR Med Educ. 2019;5(1):e13930.
6. 6. Pinto Dos Santos D, Giese D, Brodehl S, Chon SH, Staab W, Kleinert R, et al. Medical students’ attitude towards artificial intelligence: a multicentre survey. Eur Radiol. 2019;29(4):1640-6.
7. 7. Sit C, Srinivasan R, Amlani A, Muthuswamy K, Azam A, Monzon L, et al. Attitudes and perceptions of UK medical students towards artificial intelligence and radiology: a multicentre survey. Insights into imaging. 2020;11(1):14.
8. 8. Karaca O, Çalışkan SA, Demir K. Medical artificial intelligence readiness scale for medical students (MAIRS-MS)- development, validity and reliability study. BMC Med Educ. 2021;21(1).
9. 9. Hamad M, Qtaishat F, Mhairat E, Al-Qunbar A, Jaradat M, Mousa A, et al. Artificial intelligence readiness among Jordanian medical students: using medical artificial intelligence readiness scale for medical students (MAIRS-MS). J Med Educ Curri Dev. 2024;11:23821205241281648.
10. 10. Li Q, Qin Y. AI in medical education: medical student perception, curriculum recommendations and design suggestions. BMC Med Educ. 2023;23(1):852.
11. 11. Jebreen K, Radwan E, Kammoun-Rebai W, Alattar E, Radwan A, Safi W, et al. Perceptions of undergraduate medical students on artificial intelligence in medicine: mixed-methods survey study from Palestine. BMC Med Educ. 2024;24(1):507.
12. 12. Pucchio A, Rathagirishnan R, Caton N, Gariscsak PJ, Del Papa J, Nabhen JJ, et al. Exploration of exposure to artificial intelligence in undergraduate medical education: a Canadian cross-sectional mixed-methods study. BMC Med Educ. 2022;22(1):815.
13. 13. Jackson P, Ponath Sukumaran G, Babu C, Tony MC, Jack DS, Reshma VR, et al. Artificial intelligence in medical education-perception among medical students. BMC Med Educ. 2024;24(1):804.
14. 14. Allam RM, Abdelfatah D, Khalil MIM, Elsaieed MM, El Desouky ED. Medical students and house officers’ perception, attitude and potential barriers towards artificial intelligence in Egypt, cross sectional survey. BMC Med Educ. 2024;24(1):1244.
15. 15. Lugito NP, Cucunawangsih C, Suryadinata N, Kurniawan A, Wijayanto R, Sungono V, et al. Readiness, knowledge, and perception towards artificial intelligence of medical students at faculty of medicine, Pelita Harapan University, Indonesia: a cross sectional study. BMC Med Educ. 2024;24(1):1044.
16. 16. Tung AYZ, Dong LW. Malaysian medical students’ attitudes and readiness toward AI (artificial intelligence): a cross-sectional study. J Med Educ Curri Dev. 2023;10:23821205231201164.