By Ellyn Murakami (24-25)

Why is this study important?

Organisations continue to navigate volatile, uncertain, complex, and ambiguous (VUCA) challenges, such as globalisation, digitisation, and health crises, which require leaders to adapt and exhibit new behaviours to motivate and mobilise others effectively (Lawrence, 2013). However, perceptions of how well leadership development initiatives equip leaders to handle such complexities are consistently low. In 2014, McKinsey found only 7% of senior leaders believed they were developing global leaders effectively, and, similarly, Deloitte found only 13% of organisations thought they had done a quality job training their leaders (Gurdjian et al., 2014; Schwartz et al., 2014). These numbers are alarming considering that in 2018, it was estimated that organisations spent $370 billion on leadership development solutions globally with the USA alone spending $169 billion (Training Industry, 2020). Whilst there has been a surge of scientific interest in leadership development over the last 20 years, the field remains underdeveloped compared to the general leadership field (Day, 2024). Thus, this study aims to identify the enablers and barriers to effective leadership development in the 21st century.

What was found?

            17 individuals from seven industries were interviewed to explore their experiences of leadership development initiatives. To ensure diverse perspectives, participants were recruited across three roles: commissioners (clients of leadership development vendors), providers (E.g., consultants, in-house leadership development professionals, etc.), and participants of leadership development programmes. 

            Six themes were identified, representing the complexity of effective leadership development in the 21^st century (see Figure 1.) Effective leadership development works as a system, influenced by what happens before, during, and after a programme. Success also depends on having a well-designed measurement plan and on the wider organisational context.

Figure 1

Themes, subthemes, and illustrative quotes.

Note: Green text represents enablers and red text represents barriers. 

1. Leadership Development System

Participants consistently emphasised that leadership development works best when treated as a system, not just a course. This means it is crucial to consider what happens before, during, and after programme, how success is measured, and the organisation’s context. In this view, development is more of an ecosystem and ongoing process than a one-off classroom event.

2. Before a Leadership Development Programme

Common barriers to effective leadership development before a programme starts are not properly diagnosing the organisation’s strategy and leadership gaps and requiring participation. Without this, programmes feel disconnected from strategy, may not address the organisation’s needs, and participants may not know why they’re there. Also, requiring attendance also undermines motivation, which can negatively affect the whole group.

3. During a Leadership Development Programme

Three main enablers stood out during programmes:

1. In-person cohorts over time: participants shared that building trusting relationships and having sufficient time during and between sessions to learn, connect, and reflect were pivotal in making learning stick. 
2. Self-awareness: almost everyone cited this as the most impactful skill to learn and as a prerequisite for future leadership growth.
3. Experiential learning: challenging projects where participants applied new skills were seen as the most effective.

4. After a Leadership Development Programme

Many participants shared the importance of ongoing support after leaving the classroom. They noted that learning often fizzles due to not having enough time to continue applying and unsustained accountability, whether that comes from themselves, the programme, their leader, or the organisation. Participants also suggested check-ins, bite-sized follow-up workshops, or alumni meetups as ways to keep momentum going.

5. Measurement Plan

            Designing a measurement plan for leadership development initiatives is necessary to evaluate their effectiveness; however, this is rarely done and is challenging. Participants highlighted the importance of identifying key metrics and measuring a baseline, followed by immediate and longer-term evaluation to determine the lasting impact.

6. Context of the Organisation

Many participants emphasised that organisational culture and leadership behaviours strongly impact their development. Without support from managers or when leaders fail to model the behaviours participants are learning, development efforts quickly lose traction. Conversely, when senior leaders participate in and sponsor initiatives, it signals priority, builds a common language, and reinforces change.

What does this mean?

The study found that leadership development is most effective when approached as a holistic system with enablers and barriers at every stage before, during, and after programmes, and within the measurement plan and organisational context. Based on these findings, leadership development systems can be classified into four types. The programme and the organisational context are crucial because participants reported spending most of their working lives embedded in the organisational context, with only limited time devoted to formal programmes. As a result, even well-designed programmes may not compensate for an unsupportive organisational environment, just as a strong context may be insufficient to make up for a poorly designed programme. This interdependence is illustrated in Figure 2, which depicts four leadership development system types.

Figure 2

Leadership Development System Types

Becoming aware of which leadership development system type an organisation fits is the first step in identifying what needs to change to develop leaders most effectively. The development of an assessment would allow organisations to identify their type. If they are not in the optimal zone, this evaluation can highlight the weak components that need improvement and offer solutions.

What can organisations do going forward?

Organisations should approach leadership development holistically to ensure the most effective use of the time and money invested. Specifically, some recommendations are:

• Participants in leadership development systems should integrate new behaviours and leadership goals into their personal development plans to maintain priority and ensure accountability.
• Managers should support their team members’ leadership development by encouraging participation in growth opportunities, protecting time to practise new skills, and reinforcing accountability by sharing feedback.
• Managers and organisational leaders should model desired behaviours by participating in and sponsoring leadership development systems.
• Organisations should partner with Organisational Psychologists to diagnose needs and design, implement, and evaluate leadership development initiatives.

            In summary, the study found that leadership development is most effective when approached as a holistic system. This allows organisations to prioritise enablers and address barriers at every stage before, during, and after programmes, and within the measurement plan and organisational context. Viewed this way, leadership development becomes more than individual programmes: it becomes a system that helps leaders and organisations reach their full potential.

References

Day, D. V. (2024). Developing leaders and leadership: Principles, practices, and processes (First). Palgrave Macmillan Cham. https://doi.org/10.1007/978-3-031-59068-9

Gurdjian, P., Halbeisen, T., & Lane, K. (2014). Why leadership development programs fail. https://www.mckinsey.com/~/media/mckinsey/featured%20insights/leading%20in%20the%2021st%20century/why%20leadership%20development%20programs%20fail/why%20leadership%20development%20programs%20fail.pdf?shouldIndex=false

Lawrence, K. (2013). Developing leaders in a VUCA environment. http://www.execdev.unc.edu

Schwartz, J., Bersin, J., & Pelster, B. (2014). Global human capital trends 2014 – Engaging the 21st-century workforce. https://www.deloitte.co.uk/makeconnections/assets/pdf/global-human-capital-trends-2014.pdf

Training Industry, Inc. (2020). The size of the training industry. https://trainingindustry.com/wiki/learning-services-and-outsourcing/size-of-training-industry/

By Alice Wallace (24-25)

Introduction 

Burnout remains one of the biggest challenges in healthcare, with more than one in three professionals reporting symptoms during their careers (Nagarajan et al., 2024). Long hours, emotional demands, and exposure to distressing situations place heavy strain on staff – reducing clinical judgement and increasing the risk of errors (Matsuo et al., 2022). These pressures highlight the need to support the workforce, with artificial intelligence (AI) emerging as a transformative tool for improving efficiency, accuracy, and patient outcomes (Amann et al., 2020). AI is already making tangible contributions across healthcare. It has demonstrated high accuracy in diagnosing a range of conditions: hypertension, diabetes, Alzheimer’s disease, and several cancers. Beyond diagnostics, AI is advancing precision medicine by tailoring treatments to patient’s genetic and lifestyle choices (Huang et al., 2018). 

Despite these benefits, adoption of these nuanced technologies is not a linear process. Uptake depends on the people who must work with AI in demanding environments. Many healthcare staff report concerns about ethics, accountability, and risk (Warrington & Holm, 2024); these perceptions can strongly shape willingness to engage with AI (Choudhury, 2022). Dingel et al. (2024) found that higher perceptions of risk were linked to lower willingness to adopt AI-enabled decision support systems. While these concerns highlight important organisational challenges, it is important to consider the personal resources clinicians bring to their roles. Individual factors such as occupational self-efficacy (i.e., the confidence clinicians have in their ability to manage complex challenges) can shape how staff respond to new technologies (Rigotti et al., 2008). Kuper et al. (2025) found that clinicians with greater confidence in their own judgement were less likely to rely on AI when classifying skin images as benign or malignant – suggesting that confidence may reduce willingness to engage with technological support. 

Building on the current research, this study seeks to answer two key questions: (1) How does perceived risk affect clinicians’ willingness to use AI; and (2) Does occupational self-efficacy impact this relationship?

Methods

This research was conducted as part of a MSc in Occupational Psychology at City, University of London. A total of 125 participants took part, consisting of both qualified healthcare professionals (i.e., Nurses, Paramedics, and Psychiatrists) and students completing their clinical placements (i.e., Nursing, Midwifery, and Medicine). Recruitment was carried out using several online platforms: Facebook, Instagram, LinkedIn, and Reddit). Additional recruitment took place through university channels, via student and staff mailing lists. 

The study was hosted online, where participants completed three main questionnaires: risk perception (i.e., concerns about safety or errors), willingness to use AI (i.e., intentions to adopt these tools), and occupational self-efficacy (i.e., confidence in managing workplace challenges). To ensure the results were reliable, several other factors that might influence attitudes towards AI was also considered: age, gender, clinical experience, student status, risk aversion, technology literacy, and AI literacy. Including these controls helped rule out alternative explanations and provided a clearer picture of the psychological factors most relevant to AI adoption. 

Results

Approximately one-third of participants (31.2%) reported using AI in their clinical practice or placement, suggesting that widespread integration of these tools is relatively still limited. The results highlighted four key factors that influenced these individuals’ willingness to adopt AI in their practice…

• Student status: Students were significantly more willingness to use AI compared to professionals. This suggests that openness to these technologies may vary across stages of professional development. 
• AI literacy: This captures individuals’ awareness and understanding of how AI systems operate. Here, clinicians and participants with higher levels of AI literacy were more willing to adopt these tools in their practice. 
• Technology resistance: The measure incorporates a reluctance towards adopting new technologies. The findings showed that participants with higher technology resistance were less willing to engage with AI. 
• Risk perception: This reflects how individuals evaluate the potential dangers or uncertainties associated with AI. Participants who had higher risk perceptions of AI were less willing to use these tools in practice.

Discussion 

The findings of this study reinforced that the psychological and organisational factors play an important role in shaping AI adoption within healthcare settings. Risk perception emerged as an important predictor: clinicians and students who viewed AI as uncertain or unsafe were less inclined to adopt these technologies. This is consistent with wider evidence showing that in high-stakes environments such as healthcare, potential losses often carry more weight than possible gains (Kahneman & Tversky, 1979). In practice, this reflects a professional culture that prioritises patient safety and accountability (Warrington & Holm, 2024). 

Alongside risk, AI literacy emerged as the strongest influence on willingness to adopt these technologies in clinical practice. These findings extend previous work with nursing students, where Sumengen et al. (2024) demonstrated that AI literacy was an important factor influencing willingness – the present study extends these findings to a wider range of healthcare professions. Additionally, students were more willing to adopt that compared to current healthcare professionals. This suggests that openness to these technologies seems to vary at different stages of an individual’s professional development. These findings highlight the need for structured training that develops individuals’ confidence and competence in using AI, ensuring that both current and future professionals are adequately supported. 

Contrary to expectations, higher occupational self-efficacy was linked to reduced willingness to adopt AI. This finding suggests that clinicians who are more confident in their own judgement are less willing to rely on AI, as they place greater trust in their own expertise than in external technological support. While self-efficacy is usually seen as a protective resource that fosters resilience (Bandura, 2001), in this context it may inadvertently limit openness to innovation. 

Together, these findings highlight that adoption ultimately depends on the interplay of psychological resources, professional identity, and organisational culture. Addressing an individual’s perceptions of risk, enhancing AI literacy, and embedding supportive organisational practices are important here. By prioritising these factors, healthcare systems can foster the meaningful integration of AI that strengthens rather than undermining clinical practice.

Blog references

Amann, J., Blasimme, A., Vayena, E., Frey, D., & Madai, V. I. (2020). Explainability for artificial intelligence in healthcare: a multidisciplinary perspective. BMC Medical Informatics and Decision Making, 20(1). https://doi.org/10.1186/s12911-020-01332-6

Bandura, A. (2001). Social Cognitive Theory: an Agentic Perspective. Annual Review of Psychology, 52(1), 1–26. https://doi.org/10.1146/annurev.psych.52.1.1

Choudhury, A., Asan, O., & Medow, J. E. (2022). Effect of risk, expectancy, and trust on clinicians’ intent to use an artificial intelligence system — Blood Utilization Calculator. Applied Ergonomics, 101, 103708. https://doi.org/10.1016/j.apergo.2022.103708

Dingel, J., Kleine, A., Cecil, J., Sigl, A., Lermer, E., & Gaube, S. (2024). Predictors of Healthcare practitioners’ intention to use AI-Enabled Clinical Decision Support Systems (AI-CDSSS): A Meta-Analysis based on the Unified Theory of Acceptance and Use of Technology (UTAUT) (Preprint). Journal of Medical Internet Research. https://doi.org/10.2196/57224

Huang, C., Clayton, E. A., Matyunina, L. V., McDonald, L. D., Benigno, B. B., Vannberg, F., & McDonald, J. F. (2018). Machine learning predicts individual cancer patient responses to therapeutic drugs with high accuracy. Scientific Reports, 8(1). https://doi.org/10.1038/s41598-018-34753-5

Kahneman, D., & Tversky, A. (1979). Prospect Theory: An Analysis of Decision under Risk. Econometrica, 47(2), 263. https://doi.org/10.2307/1914185

Küper, A., Lodde, G. C., Livingstone, E., Schadendorf, D., & Krämer, N. (2025). Psychological Factors Influencing Appropriate Reliance on AI-enabled Clinical Decision Support Systems: an Experimental Online Study Among Dermatologists (Preprint). Journal of Medical Internet Research, 27, e58660. https://doi.org/10.2196/58660

Matsuo, T., Yoshioka, T., Okubo, R., Nagasaki, K., & Tabuchi, T. (2022). Burnout and its associated factors among healthcare workers and the general working population in Japan during the COVID-19 pandemic: a nationwide cross-sectional internet-based study. BMJ Open, 12(11), e064716. https://doi.org/10.1136/bmjopen-2022-064716

Nagarajan, R., Ramachandran, P., Dilipkumar, R., & Kaur, P. (2024). Global estimate of burnout among the public health workforce: a systematic review and meta-analysis. Human Resources for Health, 22(1). https://doi.org/10.1186/s12960-024-00917-w

Rigotti, T., Schyns, B., & Mohr, G. (2008). A short version of the Occupational Self-Efficacy Scale: Structural and Construct Validity across five countries. Journal of Career Assessment, 16(2), 238–255. https://doi.org/10.1177/1069072707305763

Sumengen, A. A., Subasi, D. O., & Cakir, G. N. (2024). Nursing students’ attitudes and literacy toward artificial intelligence: a cross-sectional study. Teaching and Learning in Nursing. https://doi.org/10.1016/j.teln.2024.10.022

Warrington, D. J., & Holm, S. (2024). Healthcare ethics and artificial intelligence: a UK doctor survey. BMJ Open,14(12), e089090. https://doi.org/10.1136/bmjopen-2024-089090