Content area
Aim
To determine the value of an artificial intelligence (AI)-image generation learning sequence on higher-education nursing student self-reflection and recognition of unconscious bias in the context of disability.
Background
Self-reflection and recognition of bias amongst undergraduate nursing students enhances reasoning skills and self-awareness in clinical situations. Teaching self-reflection to a diverse cohort can be challenging, making it essential to develop and assess innovative technological tools that support engagement in reflective practice.
Design
A multi-methods approach was adopted, obtaining both quantitative and qualitative data for analysis through a survey.
Methods
Twenty-nine nursing students from the Australian Catholic University were surveyed. Qualitative data underwent both content and inductive thematic analysis. Quantitative data were summarised using descriptive statistics. The study is reported according to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) cross-sectional study guideline.
Results
AI-image generation aided self-reflection on personal views about disability and recognition of potential personal and society biases towards disability amongst 90 % (n = 26) and 70 % of participants respectively. Visualisation of thoughts supported self-reflection and identification of generalisations held about disability. Eighty percent of respondents felt AI-image generation prompted them to consider how views and biases about disability may influence nursing practice. AI-image generation was identified to be an interesting and novel tool for self-reflection.
Conclusion
Findings suggest AI-image generation may be a useful tool in supporting students to practice self-reflection and identify unconscious biases. AI-image generation may assist students to consider how personal views can impact on clinical practice.
1 Introduction
Technology-enabled learning is the use of technologies to support student-focused learning ( Sen and Leong, 2019). Technology, including digital content, systems, platforms, portable devices and cloud computing play a vital role in the facilitation of learning in higher-education ( Al-Hariri and Al-Hattami, 2017; Forehand et al., 2017; Gause et al., 2022; Hasim, 2018; Western Sydney University, 2023). With dramatic evolution in technology, tertiary education has been called on to place ‘teaching as design’ at the centre of contemporary higher education ( Goodyear, 2015). This entails focusing on pre-active planning to design evidence-based innovative strategies and learning tasks that are supported by physical and digital environments ( Goodyear, 2015). Whilst higher education-based technology traditionally focused on delivering instruction, emerging technologies are now shaping educational practices and pedagogy, prompting the development of novel ways to engage students in creative learning experiences ( Veletsianos, 2011).
One innovative technology that seemingly crept up on higher education is generative artificial intelligence (AI). Lim et al., (2023) defines generative-AI as the use of algorithms based on existing data to create new data, such as text and images. Whilst AI challenges academic integrity ( Gulumbe et al., 2024) and causes disruption to the higher education field ( Popenici, 2023), much like the invention of the calculator in the early 1960s, AI technologies have also been found to successfully promote the student learning experience ( Hanna and Liu, 2023). Reed et al., (2023) found the use of mindtools through the form of AI generated images to be an effective teaching and learning tool, promoting students to examine their perceptions and fears of the nursing profession. Using words and students’ responses to questions/reflective prompts to generated AI-images, is a way to bring their “words and mental representations to life in visual form” ( Reed et al., 2023, pp. 373). Reed et al., (2023) encouraged further research into the use of generative AI in nursing education. One area recommended for further research was to use generative AI to promote self-reflection on specific patient encounters to promote increased personal reflection ( Reed et al., 2023).
Self-reflection amongst undergraduate nurses has been identified to enhance reasoning skills and self-awareness in clinical situations ( Bjerkvik and Hilli, 2019) and is recognised as an essential element in nursing practice ( Barchard, 2022), as it supports the integration of experiences, theory and practice into learning and development ( Barbagallo, 2021). Perry’s model (1970) identifies the need to move students towards meaning making, where knowledge is integrated with experience and reflection into practice. Through promoting self-reflection within learning materials, students are encouraged to draw on their experiences and knowledge and personal beliefs and opinions, which promotes a developmental narrative and fosters constructivist learning theory, whereby students scaffold and build on learning based on their personal and sequential learning experiences ( Angelo, 2012; Biggs and Tang, 2011).
At the end of their nursing degree, students studying a Bachelor of Nursing in Australia will be eligible for registration as a nurse with the Australian Health Practitioners Regulatory Authority (AHPRA). The Nursing and Midwifery Board AHPRA Registered Nurse Standards for Practice (2016) standard 1 requires nurses to be able to “develop their practice through reflection on experiences, knowledge, actions, feelings and beliefs to identify how these shape practice” (pg. 3). Supporting students to develop and refine their personal reflection skills through technology-based learning sequences may assist them to achieve and implement this standard of nursing practice. The aim of this research was to determine the usefulness of artificial intelligence (AI)-image generation on nursing student self-reflection and recognition of unconscious bias in the context of people with a disability.
2 Methods
A cross-sectional study of NRSG372 Principles of Nursing Chronic Illness and Disability semester 1, 2024 students at Australian Catholic University (ACU) campuses nationally who completed an AI image generation activity, was conducted to assess the impacts AI image generation has on self-reflection and recognition of unconscious bias. Whilst it is acknowledged that involving students in research presents some ethical challenges (unequal relationships, conflict of interest, obligation to participate, confidentiality concerns), incorporating solutions to these ethical challenges into the research methodology promoted collection of accurate, unbiased data. A multi-method approach was adopted, to provide opportunity for the expansion of quantitative results with qualitative data ( Greenhalgh et al., 2020).
The ontological assumption underpinning this study was relativism, whereby each participant’s reality is subjective and varied ( Parahoo, 2014) and dependent on individual interpretation ( Bradshaw et al., 2017). Understanding the topic relied on obtaining the subjective perspectives and opinions of participants ( Bradshaw et al., 2017). A constructivist view was taken in approaching the study, with the aim to learn from the experiential accounts of participants, regarding their view of AI-image generation. A grounded theory approach was used to draw themes out from the data, allowing meaning to emerge from the data, to explain the impact of AI-image generation on self-reflection and recognition of unconscious bias.
The study was reviewed and approved by the Human Research Ethics Committee at the ACU (review number 2024–3624E).
2.1 Learning sequence to be evaluated
Building on the work and recommendations of Reed et al., (2023) and the need to promote self-reflection amongst nursing students, an AI image generation learning sequence was incorporated into the ACU nursing curriculum, with the aim to determine the usefulness of artificial intelligence (AI)-image generation on nursing student self-reflection and recognition of unconscious bias in the context of people with a disability.
The learning sequence was incorporated into the unit NRSG372, a third-year subject sitting within the Bachelor of Nursing, Bachelor of Nursing/Bachelor of Paramedicine and Bachelor of Nursing/Bachelor of Business Administration, being undertaken by students who will apply for registration as a nurse with AHPRA. One of the learning outcomes of NRSG372 is for students to “reflect on the lived experiences of the person with chronic illness and/or disability.” Incorporating a technology enabled learning sequence that encourages personal reflection on beliefs/feelings/stigmas about/towards disability, aimed to support students to work towards achievement of this learning outcome. Incorporating a reflective activity into the student’s learning journey may challenge students to reflect on their personal feelings and biases, with the aim to encourage future reflection on the lived experiences of individuals with long-term conditions.
The AI image generation learning sequence incorporated into self-directed online learning modules within NRSG372 included five interrelated tasks (Supplementary File 1: Learning Sequence) that the students were encouraged to complete:
1) watching a TED talk about invisible disability.
2) writing a response to a descriptive prompt about disability.
3) using their response to inform AI-generated image development.
4) viewing the AI-generated images and answering reflective questions.
5) Participating in collaborative reflection with another student.
It was anticipated that the sequence would take approximately 30-minutes to complete. As the learning sequence was embedded into a self-directed online learning module, it was not compulsory for students to complete the task. This learning sequence represented the first time an AI-image generation learning sequence was incorporated into the ACU nursing curriculum to promote self-reflection. It is therefore important to evaluate this innovative teaching and learning strategy to assess whether it has been perceived to be valuable in promoting self-reflection and recognising unconscious bias.
2.2 Participants and eligibility
No demographic data were obtained; however, it was anticipated that most participants would be female, due to the overrepresentation of females in nursing cohorts. A response rate of between 5 % and 10 % of eligible participants was anticipated. Eligible participants were students who were enrolled in NRSG372 in semester 1, 2024 and had completed the self-directed AI-image generation activity. The learning sequence was embedded into a self-direct learning module and was not compulsory. Therefore, students did not have to complete the task. The sequence was not an assessable item, rather a way to students to personally explore their perceptions about disability. Whilst completion of the task was encouraged, engagement in online learning has been observed to be low within the unit and across nursing units of study at ACU.
2.3 Instrumentation
Data were collected via an online survey using the Qualtrics platform. An online survey was chosen for data collection to reduce student burden. Students were able to complete the survey at any time that suited them within the data collection period. Additionally, as there was no current workload or funding allocation, surveys presented a low-cost data collection method that could be conducted quickly ( Nayak and Narayan, 2019). Surveys also allowed for complete anonymity and deidentification of participants, such that would not be possible through interviews or focus groups, thus promoting respect for each participant.
A survey instrument was developed (Appendix 1), as no survey instrument about AI-image generation and self-reflection was found. The survey instrument incorporated two initial screening questions to ensure 1) participants were students in NRSG372 in semester 1, 2024 at ACU and 2) the student completed the AI-image generation activity. No demographic data were sought to ensure anonymity of participants. Additionally, students were provided with detailed participant information (plain language statement) to ensure informed consent. The plain language statement included information about alternatives to participation, rights to withdraw (how and when they can withdraw and that there was/is no penalty/detriment if they do withdraw), confidentiality and anonymity, voluntariness, ethical oversight (ethics approval information) and exactly what the research involved so students could make an informed decision regarding their participation, thus promoting respect for the value of each student ( NHMRC, 2023). Students were advised that participation was voluntary and would not impact on their grades. Further, the plain language statement included information about who was conducting the research, potential conflicts of interest, the purpose of the study and how the results will be used and disseminated, to support students to understand the merit of the study ( Office for Learning and Teaching, 2016). Student support/debrief opportunities were provided through advising students of the availability of free counselling through ACU as well as free 24-hour mental health telephone support.
Open-ended questions (qualitative data) were used to obtain feedback on the value of the AI-image generation learning sequence on student self-reflection and recognition of unconscious bias. Likert scales were used to gain insight into the perceived value of the learning sequence. The survey tool was piloted with three ACU nursing students to gain feedback on feasibility, content and face validity, as well as reviewed by three experienced academics, with minor amendments being made based on feedback received.
2.4 Data collection
Invitations to participate in the online survey were sent via an announcement on the NRSG372 online learning platform. This announcement auto generated an email to each student who was enrolled in NRSG372 in Semester 1, 2024 nationally (n = 1795). The National Lecturer in Charge (NLIC) made this announcement on the 22 nd of April 2024. The ACU online learning platform and enrolment system work together to ensure students who were enrolled in NRSG372, were added to the NRSG372 online learning platform. Therefore, the NLIC did not need to personally identify each potential participant as this information was automatically completed through the enrolments system. The online survey remained open for a period of two weeks. A reminder announcement was made by the NLIC via the NRSG372 online learning platform at the beginning of the second week (30th April 2024). This reminder autogenerated a reminder email to all students enrolled in NRSG372 in semester 1, 2024 nationally. All participants were advised that the online survey was voluntary and anonymous and had no bearing on their grades or relationship with the university/lecturers.
2.5 Data analysis
An inductive thematic analysis of qualitative data were conducted, to identify key themes. As identified by Braun and Clarke (2006)/(2021), this consists of six phases whereby themes are created, named and defined to explain and interpret the content. Examples of the identified themes are selected in the final phase (phase 6) and related back to the research aim. Following the thematic analysis of qualitative data, a separate content analysis was conducted using NVIVO15. Participant’s responses to questions were read multiple times and each response was broken down into key ideas. Each idea was coded before similar codes were grouped together into broader themes, based on shared meaning. The number of responses aligning with each theme was counted and illustrative quotes assigned to each theme. Given the exploratory nature of the study and the small number of participants who provided short-answer responses, thematic saturation was not anticipated. Instead, responses were examined to identify illustrative themes and insights that reflected participants’ perspectives. In both the thematic and content analysis, qualitative data were coded by three researchers to ensure accurate interpretation of data and applicable generation of key themes and coding of data, promoting merit and integrity of the research. Quantitative data were analysed using IBM SPSS Statistics. Descriptive statistics were used to summarise data, reporting numbers and percentages. Comparisons ( t-test/Mann Whitney U test) were not conducted as no demographic variables were obtained to create groups for comparison.
3 Results
Whilst the invitations to participate were sent to all students enrolled in NRSG372 in semester 1, 2024 (n = 1795), only students who had completed the AI-image generation activity were be eligible to complete the survey. In reviewing the NRSG372 online learning platform analytics, just 87 of the 1795 students (4.8 %) completed the non-compulsory AI-image generation learning sequence. Therefore, it was expected that 4–9 students would respond to the survey invitation (5–10 % response rate anticipated). Twenty-nine out of the eligible 87 students (33.3 %), however, met eligibility criteria (enrolled in NRSG372 semester 1, 2024 and competed the AI-image generation activity) and consented to participate in the survey. Some students did not complete all parts of the survey as aside from eligibility questions, no questions were compulsory. Therefore, within the results, the percentages reported are based on the number of students that answered the respective question. Five key themes emerged from the thematic analysis of the data.
3.1 AI-image generation promotes self-reflection
Near 90 % (n = 26) of the participants indicated that the AI-image generation activity helped them to self-reflect on their opinions, beliefs and views about disability (
“I am a visual person and by creating the AI image it enabled me to visualise what having a disability looked like and facilitated me to practice self-reflection”
“The AI image generation activity promoted self-reflection by encouraging users to visualise abstract concepts or personal experiences, prompting them to delve deeper into their thoughts and emotions.”
Participants identified that viewing the AI-images prompted them to reflect on their own definition of what disability is:
“It allowed for a visual prompt to reflect on our own definitions of disability”
“(I was) confronted with (my) own perceptions and beliefs on what someone else’s experiences are like”
Some participants found the activity supported them to consider different types of disability and the variable impacts of disability:
“It made me realise that not all disabilities are physical as all the images have someone in a wheelchair but there are various types of disabilities that I may come across as a nurse.”
“It had (a) positive influence (on my) understanding about disability. It clarified that disability may have (the) same cause but (a) different result and intensity for the patient.”
Content analysis found 10 of the 13 responses identified AI
3.2 AI-image generation promotes recognition and challenge of personal biases
Over 70 % (n = 16) of respondents indicated that the AI-image generation activity helped them recognise and visualise potential biases they may have towards disability (
Table 1). Biases were able to be visualised and reflected to students, to further promote self-reflection:
“(I) was able to visualise how my biases appeared from another perspective.”
“It allowed (me) to physically see potential biases, then after seeing physically our own views it has allowed us to reflect on our limitations in providing care if we only have a set view.”
Additionally, the AI image generation activity led some students to reflect on their personal views of/bias towards individuals with disabilities, prompting them to challenge their own perceptions:
“To visually see what my words created in a photo allowed me to reflect on how I see individuals with a disability.”
“It made me recognise that I often see people with a disability as being lonely and unsupported however, this is not the case when they have a strong support system who empowers them to reach their full potential and take pride in their disability.”
“(It is) important to consider the whole person, not view someone for their limitations but rather focus on their abilities to empower them as opposed to confine them to a specific mould that has been instilled in our own biases.”
3.3 AI-image generation promotes recognition and challenge of societal biases
Participants also reported that generated AI-imagery helped them to see and reflect on underlying biases and generalisation held a society about disability:
“It allowed me to see the underlying bias and generalisations that society has regarding individuals with disabilities. Even though many (disabilities) are invisible It (the image) would show so many individuals in wheelchairs even when I didn't include anything about a wheelchair in my image key terms.”
The reaction to the societal biases portrayed by AI-generated imagery, motivated students to challenge these biases:
“I learnt that AI has a biased impression of the word disability and that I don't agree with the interpretation of my words represented in some of the pictures.”
“It encouraged me to actively dispel stigmas I have inherited via society.”
3.4 AI-image generation promotes consideration of how beliefs and biases can impact on clinical care and people with disabilities
Near 80 % (n = 14) of responding participants felt that the activity prompted them to consider how their opinions, beliefs, views and potential biases about disability may impact on their nursing practice (
Table 1):
“ The AI image generation activity prompted me to consider how my opinions, beliefs, views and potential biases about disability may impact my nursing practice by encouraging me to reflect on how these factors could influence my interactions with patients with disabilities. It also highlighted the importance of recognising and addressing any biases to ensure equitable and respectful care for all patients.”
“Prompted reflection on my practice as a nurse to change my view from something more negative to a more positive outlook as to not impose my own beliefs on someone’s experience.”
The potential impact of a nursing student’s behaviour towards people with disabilities was considered, with respondents reflecting on how their personal approach can either support or hinder individuals with disabilities:
“The way that I behave can really impact a person with a disability, these are strong and resilient individuals. They require advocacy but also making sure that I allow them to do all they can rather than trying to help with everything.”
“It prompted (me) to think about the people I may encounter and the challenges they might have or the support they may or may not need. I realise that my own personal strength and abilities may influence my understanding of other people's attitudes and coping mechanisms.”
3.5 AI-image generation is a creative and novel way to promote self-reflection
Over 80 % (n = 13) and near 70 % (n = 11) of responding participants agreed or strongly agreed that AI-image generation is a valuable tool to support self-reflection and recognition of biases respectively (
“New creativity and innovation”
“The use of the AI question to generate an image in itself is really interesting.”
“An interesting way to explore perspective and attitudes.”
“It was a fun activity which allowed for the ability to physically see our own personal views.”
Some participants identified that the activity did not help them to self-reflect (10 %, n = 3), recognise potential biases (27 %, n = 6) or the impacts opinions, beliefs, views and potential biases about disability may have on their nursing practice (21 %, n = 4) (
Table 1). Minimal qualitative feedback was provided as to why the activity did not support self-reflection and recognition of potential biases, however, the structure, wording and way the activity was presented and interpreted by participants may have impacted on the opinion and usefulness of the activity:
“The AI image generation activity likely didn't help recognise potential biases towards disability because it focused more on visual representation and creativity rather than challenging preconceived notions or biases. Additionally, biases related to disability are complex and deeply ingrained, requiring more nuanced and intentional reflection to uncover.”
“ The question that could be asked was limited to a particular structure.”
“Because you are not able to accurately interpret bias in two sentences. Disability is a complex subject and very individual to each person or case so that makes it difficult to condense into a few words.”
Content analysis of the data identified similar overlapping themes (
4 Discussion
Incorporating both imagery and AI technologies in nursing education holds significant potential for enhancing student learning through reflection and critical self-awareness. Imagery, as a tool for reflection, has long been a part of nursing education, fostering deeper understanding and emotional engagement in students. Nicol and Pocock (2020) demonstrated this by showing how art-based workshops promoted reflective and critical discussions about death and dying amongst undergraduate nursing students. Similarly, Anderson et al. (2022), found that visual images used during simulation debriefing helped nursing students engage more deeply in self-reflection. These studies underscore the importance of visual stimuli in supporting students' ability to engage critically with emotionally and ethically complex subjects. Lapum and St-Amant (2016), further emphasised the use of visual images to prompt critical reflection, noting that such strategies increased student engagement in the reflective process even in large classroom settings. Collectively, these findings highlight how imagery serves as a powerful medium for reflection in nursing education. However, the integration of AI into this process represents a new frontier, combining traditional methods with cutting-edge technology to create even more dynamic reflective learning environments.
The use of AI across all sectors, including higher education has grown exponentially ( Chu et al., 2022), offering innovative tools that support learning, assessment and student engagement ( Chu et al., 2022). AI is now used in a range of higher educational functions, including automatic assessment, feedback generation and evaluation of student activities ( Crompton and Burke, 2023). Crompton and Burke (2023) add that AI acts as both an assistant and tutor, as well as helping to manage learning analytics and patterns, which in turn can shape personalised student learning experiences. The adoption of AI a academia is expanding rapidly; for example, in a study involving 30 teaching staff at a leading university in Australia, 48.3 % (n = 14) indicated that they were using or planning to use generative AI in the area of teaching and just one interviewee reported that they had never used AI to support their work, further highlighting the growing presence of AI in education ( Lee et al., 2024).
A nursing education, AI is also being embraced by students. In a study involving 204 nursing students, near 20 % of participants engaged with ChatGPT and 23.5 % used AI features in programs like PowerPoint as part of their learning process ( Jallad et al., 2024). The increasing incorporation of AI into educational curricula offers opportunities to revolutionise traditional teaching and learning methods, paving the way for innovative learning designs that are not only dynamic but also highly personalised (Imran et al., 2024). In the context of this study, these advancements in AI were used to explore the intersection of AI-generated imagery and self-reflection. The work of Reed (2023) laid the foundation by combining AI-generated images with reflective discussions in nursing education. Reed’s research demonstrated that AI-generated images could effectively challenge students' perceptions of professional identity and prompt them to critically reflect on cultural stereotypes in nursing. Building on this, our study encouraged nursing students to generate their own AI images using prompts related to disability. This process enabled students to confront their personal biases, visualising the often-unconscious stigmas surrounding disability. Through this reflective practice, students engaged in deeper critical thinking about the impact of their biases on both patient care and broader societal attitudes toward disability.
The role of self-reflection in nursing education is pivotal in developing clinical judgment and critical thinking ( Bulman et al., 2013). Self-reflection is not merely a passive activity but a critical tool that fosters emotional intelligence, ethical reasoning and the ability to provide compassionate, unbiased care. In this study, participants indicated that AI-image generation supported self-reflection. These findings reinforce those from Saritepeci and Durak (2024) whereby design-based learning promoted critical reflection and reflection development amongst undergraduate teaching students, reinforcing the potential of AI in nurturing reflective capabilities.
Additionally, participants in the current study indicated that AI image generation aided them to recognise and visualise personal biases about disability, assisted with the identification of societal stigmas about disability and supported them to consider the impact personal beliefs, views and biases may have on nursing practice. These findings resonate with research exploring the relationship between bias and nursing care. Grove et al. (2021) identified that personal biases, especially those related to characteristics such as race, gender, or disability, can negatively impact the quality of care delivered by nurses. Gatewood et al. (2019) also explored this concept through an online and face-to-face implicit bias activity in nursing education, where students reported finding the activities helpful in increasing awareness of their biases and acknowledged that recognising these biases could help improve their future nursing clinical practice. Activities involving AI-generated imagery can act as catalysts for students to confront and deconstruct their unconscious biases and thus impact on their care delivery. Encouraging students to reflect on, consider and recognise their own feelings and unconscious bias may assist them to learn about the influence of unconscious bias on health disparities, their nursing practice and the importance of managing personal bias ( Schultz and Baker, 2017).
Through this study, it was found that using AI to generate personalised visual representations of disability allowed nursing students to externalise and critically reflect on their unconscious assumptions about disability. This reflective process encouraged them to consider how these biases might shape and influence their nursing practice and, potentially lead them to reflect on how they can actively manage such biases to provide more inclusive care. AI-image generation, as a novel tool in nursing education, not only enhances self-reflection but also supports the critical task of bias recognition. By allowing students to self-create and reflect on AI-generated images that visualise their thoughts and feelings, educators can deepen students’ understanding of their unconscious biases, ultimately promoting more equitable and compassionate nursing practice. It is, however, important to note that the qualitative findings presented in this study are not intended to represent a saturated or comprehensive set of themes. Rather, they provide preliminary insights into the value of AI-image generation in self-reflection and recognition of unconscious bias. While limited by the small number of responses, these findings highlight areas of interest that warrant further exploration in larger, dedicated qualitative studies where thematic saturation could be more fully achieved. Future research should continue to explore the intersection of AI, imagery and reflective practice, particularly in relation to addressing other forms of bias in healthcare.
Another consideration for future research may include analysing engagement in self-directed online learning activity. Through this research it was identified that very few students (less than 5 %) completed the AI-image generation activity and associated online-learning module, as it was not compulsory and not directly linked to an assessment task. Additional research should be undertaken to understand online-learning engagement and participation rates and to assess the value (or not) of such resources.
5 Limitations
As the AI-image generation learning sequence was embedded within a self-directed online learning module, it was not compulsory for students to complete the activity. This provides explanation as to the low participation rates in the learning sequence and research surveys. Student engagement in the online-learning modules within NRSG372 and nursing units at ACU has been observed to be low. Additional research into engagement in online-learning within the ACU nursing curriculum is planned for late 2025. The sample size is acknowledged as a limitation of the study and as such findings cannot be generalised for the population. Sampling bias may have occurred as nursing students who were interested in the topic or engaged in their learning may have been more likely to participate in the study. Further, use of a non-validated survey tool decreases the validity and reliability of the findings. The questionnaire was, however, piloted with three ACU nursing students to gain feedback on feasibility, content and face validity, as well as reviewed by three experienced academics, with minor amendments being made based on feedback received.
Additionally, while the study provides important insights into the value of AI-image generation in promoting self-reflection and recognition of unconscious bias amongst nursing students, the findings should be interpreted with caution due to missing data. Although 29 participants consented, only 16 completed the entire survey and some items had up to 50 % missing responses. As no questions were compulsory, participants were free to skip any items, which likely contributed to incomplete responses, particularly for short-answer questions that required more time or detail. This pattern of missing data may have limited the richness of qualitative findings and introduced potential bias, as it is possible that those who chose not to respond may have held different views. Despite this limitation, the data that were available still provide valuable preliminary insights into the usefulness of a novel learning sequence using AI-image generation. Future studies with larger samples and strategies to minimise item non-response (e.g., streamlined survey design or partial mandatory fields) would strengthen the evidence base.
6 Conclusion
This study provides insight into the value of using AI-image generation to support nursing student self-reflection and recognition of unconscious bias in the context of disability. Participants indicated that AI-image generation is a valuable tool to promote self-reflection, recognise unconscious biases, identify social stigmas and consider the impact of opinions, beliefs, views and potential biases about disability on nursing practice. Self-reflection and recognition of bias is fundamental to enhancing reasoning skills in undergraduate nurses. Further, nurses need to develop their practice through reflection on experiences, knowledge, actions, feelings and beliefs to identify how these shape practice. Using innovative technological tools, such as AI-image generation may promote self-reflection and support recognition of unconscious bias in nursing students and thus positively challenge their approach to patient care in the context of disability.
Ethical approval
The study has been reviewed and approved by the Human Research Ethics Committee at the Australian Catholic University (review number 2024–3624E).
Funding statement
The authors received no financial support for the research, authorship, and/or publication of this article.
Funding statement
The authors received no financial support for the research, authorship and/or publication of this article.
CRediT authorship contribution statement
Wendy Luck: Writing – review & editing, Formal analysis. Bethany Arbuckle: Writing – review & editing, Formal analysis, Conceptualization. MULLAN LEANNE: Writing – review & editing, Writing – original draft, Visualization, Project administration, Methodology, Investigation, Formal analysis, Conceptualization.
Informed consent statement
All participants were provided with a detailed participant information sheet outlining the research, process for consent and withdrawal. An electronic declaration of consent was obtained from each participant, prior to completion of the online survey.
Declaration of Competing Interest
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article
Acknowledgements
We would like to acknowledge the students who gave their time to participate in this research.
Appendix A Supporting information
Supplementary data associated with this article can be found in the online version at
Appendix A Supplementary material
Supplementary material
Supplementary material
Table 1
| Yes
% (n) |
No
% (n) |
Totals
% (n) | |
| Did the AI-image generation activity help you to self-reflect on your opinions, beliefs, and views about disability? | 89.7 (26) | 10.3 (3) | 100 (29) |
| Did the AI-image generation activity help you to recognise potential biases you may have towards disability? | 72.7 (16) | 27.3 (6) | 100 (22) |
| Did the AI-image generation activity prompt you to consider how your opinions, beliefs, views, and potential biases about disability may impact on your nursing practice. | 78.9 (14) | 21.1 (4) | 100 (18) |
Table 2
| Strongly Agree / Agree
% (n) |
Undecided
% (n) |
Strongly Disagree / Disagree
% (n) |
Total
% (n) | |
| The AI-image generation activity helped me to reflect on my opinions, beliefs, and views about disability. | 82.4
(14) |
11.8
(2) |
5.9
(1) |
100 % (17) |
| The AI-image generation activity helped me to recognise potential biases I may have towards disability. | 75.0 % (12) | 12.5 %
(2) |
12.5 %
(2) |
100 % (16) |
| The AI-image generation activity prompted me to consider how my opinions, beliefs, views, and potential biases about disability may impact on my nursing practice. | 81.3 % (13) | 6.3 %
(1) |
12.5 %
(2) |
100 % (16) |
| AI-image generation is a valuable tool to support self-reflection. | 81.3 % (13) | 18.8 %
(3) |
0.0 %
(0) |
100 % (16) |
| AI-image generation is a valuable tool to support recognition of biases. | 68.8 % (11) | 25.0 %
(4) |
6.3 %
(1) |
100 % (16) |
Table 3
| Question: Describe in two sentences how the AI image generation activity promoted self-reflection? | |||
| Theme | Description | Frequency (n = 13) | Example quotes |
| Biases | Recognition of personal and societal biases in AI representations and/or reflection on the impacts of such biases | 6 | “It allowed me to see the underlying bias and generalisations that society has…” |
| Reflection on nursing role (professional practice) | Insight into how beliefs influence the nurse’s role and patient care | 4 | “Prompted reflection on my practice as a nurse to change my view…” |
| Use/value of visual stimuli to prompt reflection | Reflections triggered by visualisation rather than verbal or written prompts | 5 | “It allowed for a visual prompt to reflect on our own definitions of disability” |
| Challenge preconceived notions | Realisation that AI outcomes did not match participants’ views, prompting reflection | 4 | “…it is not how I perceive disability.” |
| Self-awareness and identity | Reflection on personal values, beliefs, or societal conditioning | 4 | “it encouraged me to actively dispel stigmas I have inherited via society” |
| Reflection through creativity, innovation, engagement and enjoyment | Appreciation of the activity as an interesting, creative, novel, engaging, fun experience | 5 | “new creativity and innovation…”
“it was fun to complete” |
| Question: Describe in two sentences how the AI image generation activity helped you to recognise potential biases you may have towards disability? | |||
| Theme | Description | Frequency (n = 6) | Example quotes |
| Biases | Recognition of personal and societal biases in AI representations and/or reflection on the impacts of such biases | 5 | “…more bias that technology and society may have”
“see individuals with disability as lonely and unsupported,” “how my biases appeared.” |
| Self-awareness and reflection | Reflection on personal values, beliefs, or societal conditioning | 3 | “seeing what my words created in a photo allowed me to reflect.” |
| Question: Describe in two sentences how the AI image generation activity prompted you to consider how your opinions, beliefs, views, and potential biases about disability may impact on your nursing practice? | |||
| Theme | Description | Frequency (n = 7) | Example quotes |
| Reflection on nursing role (professional practice) | Insight into how beliefs influence the nurse’s role and patient care | 6 | “these [people with a disability] are strong and resilient individuals,”
“how my own beliefs could influence interactions.” |
| Advocacy and person-centredness | Reflection on the importance of advocacy and patient-centred care | 3 | “not view someone for their limitations,” “empower them.” |
| Biases | Recognition of personal and societal biases in AI representations and/or reflection on the impacts of such biases | 2 | “reflect on our limitations in providing care if we only have a set view.” |
| Question 4: What did you learn by completing the AI-image generation activity? | |||
| Theme | Description | Frequency (n = 9) | Example quotes |
| Biases | Recognition of personal, societal or internal biases in AI representations and/or reflection on the impacts of such biases | 5 | “AI has a biased impression of the word disability.” |
| Use/value of visual stimuli to prompt reflection | Reflections triggered by visualisation rather than verbal or written prompts | 4 | “see our own personal views,” “how I view disability visually.” |
| Reflection through creativity, innovation, engagement and enjoyment | Appreciation of the activity as an interesting, creative, novel, engaging, fun experience | 3 | “interesting way to explore perspectives.”
“It was a fun activity…” |
| Cross-Question Themes | |||
| Theme | Description | Frequency (n = 35) | |
| Biases | Recognition of personal and societal biases in AI representations and/or reflection on the impacts of such biases | 18 | |
| Reflection on nursing role (professional practice) | Insight into how beliefs influence the nurse’s role and patient care | 10 | |
| Reflection through creativity, innovation, engagement and enjoyment | Appreciation of the activity as an interesting, creative, novel, engaging, fun experience | 14 | |
| Use/value of visual stimuli to prompt reflection | Reflections triggered by visualisation rather than verbal or written prompts | 9 | |
| Self-awareness and reflection | Reflection on personal values, beliefs, or societal conditioning | 7 |
© 2025 The Authors